Convert Color To Black & White Portrait Using CSS | How To Add Filter Effects to Images

In this video will write a Css class that will convert a colored image to black and white, the grayscale image will be generated on the client side by the browser’s Css engine instead of using javascript or an image editor to upload two versions of the same picture.

There are plenty of different ways to convert images to black and white, but using Css is the most appropriate approach because layout, colors, textures, shadows…

#css

What is GEEK

Buddha Community

How to Add Splash Screen in Android and iOS with Flutter

When your app is opened, there is a brief time while the native app loads Flutter. By default, during this time, the native app displays a white splash screen. This package automatically generates iOS, Android, and Web-native code for customizing this native splash screen background color and splash image. Supports dark mode, full screen, and platform-specific options.

What's New

[BETA] Support for flavors is in beta. Currently only Android and iOS are supported. See instructions below.

You can now keep the splash screen up while your app initializes! No need for a secondary splash screen anymore. Just use the preserve and remove methods together to remove the splash screen after your initialization is complete. See details below.

Usage

Would you prefer a video tutorial instead? Check out Johannes Milke's tutorial.

First, add flutter_native_splash as a dependency in your pubspec.yaml file.

dependencies:
  flutter_native_splash: ^2.2.19

Don't forget to flutter pub get.

1. Setting the splash screen

 

Customize the following settings and add to your project's pubspec.yaml file or place in a new file in your root project folder named flutter_native_splash.yaml.

flutter_native_splash:
  # This package generates native code to customize Flutter's default white native splash screen
  # with background color and splash image.
  # Customize the parameters below, and run the following command in the terminal:
  # flutter pub run flutter_native_splash:create
  # To restore Flutter's default white splash screen, run the following command in the terminal:
  # flutter pub run flutter_native_splash:remove

  # color or background_image is the only required parameter.  Use color to set the background
  # of your splash screen to a solid color.  Use background_image to set the background of your
  # splash screen to a png image.  This is useful for gradients. The image will be stretch to the
  # size of the app. Only one parameter can be used, color and background_image cannot both be set.
  color: "#42a5f5"
  #background_image: "assets/background.png"

  # Optional parameters are listed below.  To enable a parameter, uncomment the line by removing
  # the leading # character.

  # The image parameter allows you to specify an image used in the splash screen.  It must be a
  # png file and should be sized for 4x pixel density.
  #image: assets/splash.png

  # The branding property allows you to specify an image used as branding in the splash screen.
  # It must be a png file. It is supported for Android, iOS and the Web.  For Android 12,
  # see the Android 12 section below.
  #branding: assets/dart.png

  # To position the branding image at the bottom of the screen you can use bottom, bottomRight,
  # and bottomLeft. The default values is bottom if not specified or specified something else.
  #branding_mode: bottom

  # The color_dark, background_image_dark, image_dark, branding_dark are parameters that set the background
  # and image when the device is in dark mode. If they are not specified, the app will use the
  # parameters from above. If the image_dark parameter is specified, color_dark or
  # background_image_dark must be specified.  color_dark and background_image_dark cannot both be
  # set.
  #color_dark: "#042a49"
  #background_image_dark: "assets/dark-background.png"
  #image_dark: assets/splash-invert.png
  #branding_dark: assets/dart_dark.png

  # Android 12 handles the splash screen differently than previous versions.  Please visit
  # https://developer.android.com/guide/topics/ui/splash-screen
  # Following are Android 12 specific parameter.
  android_12:
    # The image parameter sets the splash screen icon image.  If this parameter is not specified,
    # the app's launcher icon will be used instead.
    # Please note that the splash screen will be clipped to a circle on the center of the screen.
    # App icon with an icon background: This should be 960×960 pixels, and fit within a circle
    # 640 pixels in diameter.
    # App icon without an icon background: This should be 1152×1152 pixels, and fit within a circle
    # 768 pixels in diameter.
    #image: assets/android12splash.png

    # Splash screen background color.
    #color: "#42a5f5"

    # App icon background color.
    #icon_background_color: "#111111"

    # The branding property allows you to specify an image used as branding in the splash screen.
    #branding: assets/dart.png

    # The image_dark, color_dark, icon_background_color_dark, and branding_dark set values that
    # apply when the device is in dark mode. If they are not specified, the app will use the
    # parameters from above.
    #image_dark: assets/android12splash-invert.png
    #color_dark: "#042a49"
    #icon_background_color_dark: "#eeeeee"

  # The android, ios and web parameters can be used to disable generating a splash screen on a given
  # platform.
  #android: false
  #ios: false
  #web: false

  # Platform specific images can be specified with the following parameters, which will override
  # the respective parameter.  You may specify all, selected, or none of these parameters:
  #color_android: "#42a5f5"
  #color_dark_android: "#042a49"
  #color_ios: "#42a5f5"
  #color_dark_ios: "#042a49"
  #color_web: "#42a5f5"
  #color_dark_web: "#042a49"
  #image_android: assets/splash-android.png
  #image_dark_android: assets/splash-invert-android.png
  #image_ios: assets/splash-ios.png
  #image_dark_ios: assets/splash-invert-ios.png
  #image_web: assets/splash-web.png
  #image_dark_web: assets/splash-invert-web.png
  #background_image_android: "assets/background-android.png"
  #background_image_dark_android: "assets/dark-background-android.png"
  #background_image_ios: "assets/background-ios.png"
  #background_image_dark_ios: "assets/dark-background-ios.png"
  #background_image_web: "assets/background-web.png"
  #background_image_dark_web: "assets/dark-background-web.png"
  #branding_android: assets/brand-android.png
  #branding_dark_android: assets/dart_dark-android.png
  #branding_ios: assets/brand-ios.png
  #branding_dark_ios: assets/dart_dark-ios.png

  # The position of the splash image can be set with android_gravity, ios_content_mode, and
  # web_image_mode parameters.  All default to center.
  #
  # android_gravity can be one of the following Android Gravity (see
  # https://developer.android.com/reference/android/view/Gravity): bottom, center,
  # center_horizontal, center_vertical, clip_horizontal, clip_vertical, end, fill, fill_horizontal,
  # fill_vertical, left, right, start, or top.
  #android_gravity: center
  #
  # ios_content_mode can be one of the following iOS UIView.ContentMode (see
  # https://developer.apple.com/documentation/uikit/uiview/contentmode): scaleToFill,
  # scaleAspectFit, scaleAspectFill, center, top, bottom, left, right, topLeft, topRight,
  # bottomLeft, or bottomRight.
  #ios_content_mode: center
  #
  # web_image_mode can be one of the following modes: center, contain, stretch, and cover.
  #web_image_mode: center

  # The screen orientation can be set in Android with the android_screen_orientation parameter.
  # Valid parameters can be found here:
  # https://developer.android.com/guide/topics/manifest/activity-element#screen
  #android_screen_orientation: sensorLandscape

  # To hide the notification bar, use the fullscreen parameter.  Has no effect in web since web
  # has no notification bar.  Defaults to false.
  # NOTE: Unlike Android, iOS will not automatically show the notification bar when the app loads.
  #       To show the notification bar, add the following code to your Flutter app:
  #       WidgetsFlutterBinding.ensureInitialized();
  #       SystemChrome.setEnabledSystemUIOverlays([SystemUiOverlay.bottom, SystemUiOverlay.top]);
  #fullscreen: true

  # If you have changed the name(s) of your info.plist file(s), you can specify the filename(s)
  # with the info_plist_files parameter.  Remove only the # characters in the three lines below,
  # do not remove any spaces:
  #info_plist_files:
  #  - 'ios/Runner/Info-Debug.plist'
  #  - 'ios/Runner/Info-Release.plist'

2. Run the package

After adding your settings, run the following command in the terminal:

flutter pub run flutter_native_splash:create

When the package finishes running, your splash screen is ready.

To specify the YAML file location just add --path with the command in the terminal:

flutter pub run flutter_native_splash:create --path=path/to/my/file.yaml

3. Set up app initialization (optional)

By default, the splash screen will be removed when Flutter has drawn the first frame. If you would like the splash screen to remain while your app initializes, you can use the preserve() and remove() methods together. Pass the preserve() method the value returned from WidgetsFlutterBinding.ensureInitialized() to keep the splash on screen. Later, when your app has initialized, make a call to remove() to remove the splash screen.

import 'package:flutter_native_splash/flutter_native_splash.dart';

void main() {
  WidgetsBinding widgetsBinding = WidgetsFlutterBinding.ensureInitialized();
  FlutterNativeSplash.preserve(widgetsBinding: widgetsBinding);
  runApp(const MyApp());
}

// whenever your initialization is completed, remove the splash screen:
    FlutterNativeSplash.remove();

NOTE: If you do not need to use the preserve() and remove() methods, you can place the flutter_native_splash dependency in the dev_dependencies section of pubspec.yaml.

4. Support the package (optional)

If you find this package useful, you can support it for free by giving it a thumbs up at the top of this page. Here's another option to support the package:

Android 12+ Support

Android 12 has a new method of adding splash screens, which consists of a window background, icon, and the icon background. Note that a background image is not supported.

Be aware of the following considerations regarding these elements:

1: image parameter. By default, the launcher icon is used:

• App icon without an icon background, as shown on the left: This should be 1152×1152 pixels, and fit within a circle 768 pixels in diameter.
• App icon with an icon background, as shown on the right: This should be 960×960 pixels, and fit within a circle 640 pixels in diameter.

2: icon_background_color is optional, and is useful if you need more contrast between the icon and the window background.

3: One-third of the foreground is masked.

4: color the window background consists of a single opaque color.

PLEASE NOTE: The splash screen may not appear when you launch the app from Android Studio on API 31. However, it should appear when you launch by clicking on the launch icon in Android. This seems to be resolved in API 32+.

PLEASE NOTE: There are a number of reports that non-Google launchers do not display the launch image correctly. If the launch image does not display correctly, please try the Google launcher to confirm that this package is working.

PLEASE NOTE: The splash screen does not appear when you launch the app from a notification. Apparently this is the intended behavior on Android 12: core-splashscreen Icon not shown when cold launched from notification.

Flavor Support

If you have a project setup that contains multiple flavors or environments, and you created more than one flavor this would be a feature for you.

Instead of maintaining multiple files and copy/pasting images, you can now, using this tool, create different splash screens for different environments.

Pre-requirements

In order to use the new feature, and generate the desired splash images for you app, a couple of changes are required.

If you want to generate just one flavor and one file you would use either options as described in Step 1. But in order to setup the flavors, you will then be required to move all your setup values to the flutter_native_splash.yaml file, but with a prefix.

Let's assume for the rest of the setup that you have 3 different flavors, Production, Acceptance, Development.

First this you will need to do is to create a different setup file for all 3 flavors with a suffix like so:

flutter_native_splash-production.yaml
flutter_native_splash-acceptance.yaml
flutter_native_splash-development.yaml

You would setup those 3 files the same way as you would the one, but with different assets depending on which environment you would be generating. For example (Note: these are just examples, you can use whatever setup you need for your project that is already supported by the package):

# flutter_native_splash-development.yaml
flutter_native_splash:
  color: "#ffffff"
  image: assets/logo-development.png
  branding: assets/branding-development.png
  color_dark: "#121212"
  image_dark: assets/logo-development.png
  branding_dark: assets/branding-development.png

  android_12:
    image: assets/logo-development.png
    icon_background_color: "#ffffff"
    image_dark: assets/logo-development.png
    icon_background_color_dark: "#121212"

  web: false

# flutter_native_splash-acceptance.yaml
flutter_native_splash:
  color: "#ffffff"
  image: assets/logo-acceptance.png
  branding: assets/branding-acceptance.png
  color_dark: "#121212"
  image_dark: assets/logo-acceptance.png
  branding_dark: assets/branding-acceptance.png

  android_12:
    image: assets/logo-acceptance.png
    icon_background_color: "#ffffff"
    image_dark: assets/logo-acceptance.png
    icon_background_color_dark: "#121212"

  web: false

# flutter_native_splash-production.yaml
flutter_native_splash:
  color: "#ffffff"
  image: assets/logo-production.png
  branding: assets/branding-production.png
  color_dark: "#121212"
  image_dark: assets/logo-production.png
  branding_dark: assets/branding-production.png

  android_12:
    image: assets/logo-production.png
    icon_background_color: "#ffffff"
    image_dark: assets/logo-production.png
    icon_background_color_dark: "#121212"

  web: false

Great, now comes the fun part running the new command!

The new command is:

# If you have a flavor called production you would do this:
flutter pub run flutter_native_splash:create --flavor production

# For a flavor with a name staging you would provide it's name like so:
flutter pub run flutter_native_splash:create --flavor staging

# And if you have a local version for devs you could do that:
flutter pub run flutter_native_splash:create --flavor development

Android setup

You're done! No, really, Android doesn't need any additional setup.

Note: If it didn't work, please make sure that your flavors are named the same as your config files, otherwise the setup will not work.

iOS setup

iOS is a bit tricky, so hang tight, it might look scary but most of the steps are just a single click, explained as much as possible to lower the possibility of mistakes.

When you run the new command, you will need to open xCode and follow the steps bellow:

Assumption

• In order for this setup to work, you would already have 3 different schemes setup; production, acceptance and development.

Preparation

• Open the iOS Flutter project in Xcode (open the Runner.xcworkspace)
• Find the newly created Storyboard files at the same location where the original is {project root}/ios/Runner/Base.lproj
• Select all of them and drag and drop into Xcode, directly to the left hand side where the current LaunchScreen.storyboard is located already
• After you drop your files there Xcode will ask you to link them, make sure you select 'Copy if needed'
• This part is done, you have linked the newly created storyboards in your project.

xCode

Xcode still doesn't know how to use them, so we need to specify for all the current flavors (schemes) which file to use and to use that value inside the Info.plist file.

• Open the iOS Flutter project in Xcode (open the Runner.xcworkspace)
• Click the Runner project in the top left corner (usually the first item in the list)
• In the middle part of the screen, on the left side, select the Runner target
• On the top part of the screen select Build Settings
• Make sure that 'All' and 'Combined' are selected
• Next to 'Combine' you have a '+' button, press it and select 'Add User-Defined Setting'
• Once you do that Xcode will create a new variable for you to name. Suggestion is to name it LAUNCH_SCREEN_STORYBOARD
• Once you do that, you will have the option to define a specific name for each flavor (scheme) that you have defined in the project. Make sure that you input the exact name of the LaunchScreen.storyboard that was created by this tool
  • Example: If you have a flavor Development, there is a Storyboard created name LaunchScreenDevelopment.storyboard, please add that name (without the storyboard part) to the variable value next to the flavor value
• After you finish with that, you need to update Info.plist file to link the newly created variable so that it's used correctly
• Open the Info.plist file
• Find the entry called 'Launch screen interface file base name'
• The default value is 'LaunchScreen', change that to the variable name that you create previously. If you follow these steps exactly, it would be LAUNCH_SCREEN_STORYBOARD, so input this $(LAUNCH_SCREEN_STORYBOARD)
• And your done!

Congrats you finished your setup for multiple flavors,

FAQs

I got the error "A splash screen was provided to Flutter, but this is deprecated."

This message is not related to this package but is related to a change in how Flutter handles splash screens in Flutter 2.5. It is caused by having the following code in your android/app/src/main/AndroidManifest.xml, which was included by default in previous versions of Flutter:

<meta-data
 android:name="io.flutter.embedding.android.SplashScreenDrawable"
 android:resource="@drawable/launch_background"
 />

The solution is to remove the above code. Note that this will also remove the fade effect between the native splash screen and your app.

Are animations/lottie/GIF images supported?

Not at this time. PRs are always welcome!

I got the error AAPT: error: style attribute 'android:attr/windowSplashScreenBackground' not found

This attribute is only found in Android 12, so if you are getting this error, it means your project is not fully set up for Android 12. Did you update your app's build configuration?

I see a flash of the wrong splash screen on iOS

This is caused by an iOS splash caching bug, which can be solved by uninstalling your app, powering off your device, power back on, and then try reinstalling.

I see a white screen between splash screen and app

1. It may be caused by an iOS splash caching bug, which can be solved by uninstalling your app, powering off your device, power back on, and then try reinstalling.
2. It may be caused by the delay due to initialization in your app. To solve this, put any initialization code in the removeAfter method.

Can I base light/dark mode on app settings?

No. This package creates a splash screen that is displayed before Flutter is loaded. Because of this, when the splash screen loads, internal app settings are not available to the splash screen. Unfortunately, this means that it is impossible to control light/dark settings of the splash from app settings.

Notes

If the splash screen was not updated correctly on iOS or if you experience a white screen before the splash screen, run flutter clean and recompile your app. If that does not solve the problem, delete your app, power down the device, power up the device, install and launch the app as per this StackOverflow thread.

This package modifies launch_background.xml and styles.xml files on Android, LaunchScreen.storyboard and Info.plist on iOS, and index.html on Web. If you have modified these files manually, this plugin may not work properly. Please open an issue if you find any bugs.

How it works

Android

• Your splash image will be resized to mdpi, hdpi, xhdpi, xxhdpi and xxxhdpi drawables.
• An <item> tag containing a <bitmap> for your splash image drawable will be added in launch_background.xml
• Background color will be added in colors.xml and referenced in launch_background.xml.
• Code for full screen mode toggle will be added in styles.xml.
• Dark mode variants are placed in drawable-night, values-night, etc. resource folders.

iOS

• Your splash image will be resized to @3x and @2x images.
• Color and image properties will be inserted in LaunchScreen.storyboard.
• The background color is implemented by using a single-pixel png file and stretching it to fit the screen.
• Code for hidden status bar toggle will be added in Info.plist.

Web

• A web/splash folder will be created for splash screen images and CSS files.
• Your splash image will be resized to 1x, 2x, 3x, and 4x sizes and placed in web/splash/img.
• The splash style sheet will be added to the app's web/index.html, as well as the HTML for the splash pictures.

Acknowledgments

This package was originally created by Henrique Arthur and it is currently maintained by Jon Hanson.

Bugs or Requests

If you encounter any problems feel free to open an issue. If you feel the library is missing a feature, please raise a ticket. Pull request are also welcome.

---

Use this package as a library

Depend on it

Run this command:

With Flutter:

 $ flutter pub add flutter_native_splash

This will add a line like this to your package's pubspec.yaml (and run an implicit flutter pub get):

dependencies:
  flutter_native_splash: ^2.2.19

Alternatively, your editor might support flutter pub get. Check the docs for your editor to learn more.

Import it

Now in your Dart code, you can use:

import 'package:flutter_native_splash/flutter_native_splash.dart';

---

example/lib/main.dart

import 'package:flutter/material.dart';
import 'package:flutter_native_splash/flutter_native_splash.dart';

void main() {
  WidgetsBinding widgetsBinding = WidgetsFlutterBinding.ensureInitialized();
  FlutterNativeSplash.preserve(widgetsBinding: widgetsBinding);
  runApp(const MyApp());
}

class MyApp extends StatelessWidget {
  const MyApp({super.key});

  // This widget is the root of your application.
  @override
  Widget build(BuildContext context) {
    return MaterialApp(
      title: 'Flutter Demo',
      theme: ThemeData(
        // This is the theme of your application.
        //
        // Try running your application with "flutter run". You'll see the
        // application has a blue toolbar. Then, without quitting the app, try
        // changing the primarySwatch below to Colors.green and then invoke
        // "hot reload" (press "r" in the console where you ran "flutter run",
        // or simply save your changes to "hot reload" in a Flutter IDE).
        // Notice that the counter didn't reset back to zero; the application
        // is not restarted.
        primarySwatch: Colors.blue,
      ),
      home: const MyHomePage(title: 'Flutter Demo Home Page'),
    );
  }
}

class MyHomePage extends StatefulWidget {
  const MyHomePage({super.key, required this.title});

  // This widget is the home page of your application. It is stateful, meaning
  // that it has a State object (defined below) that contains fields that affect
  // how it looks.

  // This class is the configuration for the state. It holds the values (in this
  // case the title) provided by the parent (in this case the App widget) and
  // used by the build method of the State. Fields in a Widget subclass are
  // always marked "final".

  final String title;

  @override
  State<MyHomePage> createState() => _MyHomePageState();
}

class _MyHomePageState extends State<MyHomePage> {
  int _counter = 0;

  void _incrementCounter() {
    setState(() {
      // This call to setState tells the Flutter framework that something has
      // changed in this State, which causes it to rerun the build method below
      // so that the display can reflect the updated values. If we changed
      // _counter without calling setState(), then the build method would not be
      // called again, and so nothing would appear to happen.
      _counter++;
    });
  }

  @override
  void initState() {
    super.initState();
    initialization();
  }

  void initialization() async {
    // This is where you can initialize the resources needed by your app while
    // the splash screen is displayed.  Remove the following example because
    // delaying the user experience is a bad design practice!
    // ignore_for_file: avoid_print
    print('ready in 3...');
    await Future.delayed(const Duration(seconds: 1));
    print('ready in 2...');
    await Future.delayed(const Duration(seconds: 1));
    print('ready in 1...');
    await Future.delayed(const Duration(seconds: 1));
    print('go!');
    FlutterNativeSplash.remove();
  }

  @override
  Widget build(BuildContext context) {
    // This method is rerun every time setState is called, for instance as done
    // by the _incrementCounter method above.
    //
    // The Flutter framework has been optimized to make rerunning build methods
    // fast, so that you can just rebuild anything that needs updating rather
    // than having to individually change instances of widgets.
    return Scaffold(
      appBar: AppBar(
        // Here we take the value from the MyHomePage object that was created by
        // the App.build method, and use it to set our appbar title.
        title: Text(widget.title),
      ),
      body: Center(
        // Center is a layout widget. It takes a single child and positions it
        // in the middle of the parent.
        child: Column(
          // Column is also a layout widget. It takes a list of children and
          // arranges them vertically. By default, it sizes itself to fit its
          // children horizontally, and tries to be as tall as its parent.
          //
          // Invoke "debug painting" (press "p" in the console, choose the
          // "Toggle Debug Paint" action from the Flutter Inspector in Android
          // Studio, or the "Toggle Debug Paint" command in Visual Studio Code)
          // to see the wireframe for each widget.
          //
          // Column has various properties to control how it sizes itself and
          // how it positions its children. Here we use mainAxisAlignment to
          // center the children vertically; the main axis here is the vertical
          // axis because Columns are vertical (the cross axis would be
          // horizontal).
          mainAxisAlignment: MainAxisAlignment.center,
          children: <Widget>[
            const Text(
              'You have pushed the button this many times:',
            ),
            Text(
              '$_counter',
              style: Theme.of(context).textTheme.headlineMedium,
            ),
          ],
        ),
      ),
      floatingActionButton: FloatingActionButton(
        onPressed: _incrementCounter,
        tooltip: 'Increment',
        child: const Icon(Icons.add),
      ), // This trailing comma makes auto-formatting nicer for build methods.
    );
  }
}

Download Details:
 

Author: jonbhanson
Download Link: Download The Source Code
Official Website: https://github.com/jonbhanson/flutter_native_splash 
License: MIT license

#flutter #ios #android 

Top 14 Ways to Filter Pandas Dataframes Easily

Whenever we work with data of any sort, we need a clear picture of the kind of data that we are dealing with. For most of the data out there, which may contain thousands or even millions of entries with a wide variety of information, it’s really impossible to make sense of that data without any tool to present the data in a short and readable format.

Most of the time we need to go through the data, manipulate it, and visualize it for getting insights. Well, there is a great library which goes by the name pandas which provides us with that capability. The most frequent Data manipulation operation is Data Filtering. It is very similar to the WHERE clause in SQL or you must have used a filter in MS Excel for selecting specific rows based on some conditions.

pandas is a powerful, flexible and open source data analysis/manipulation tool which is essentially a python package that provides speed, flexibility and expressive data structures crafted to work with “relational” or “labelled” data in an intuitive and easy manner. It is one of the most popular libraries to perform real-world data analysis in Python.

pandas is built on top of the NumPy library which aims to integrate well with the scientific computing environment and numerous other 3rd party libraries. It has two primary data structures namely Series (1D) and Dataframes(2D), which in most real-world use cases is the type of data that is being dealt with in many sectors of finance, scientific computing, engineering and statistics.

Let’s Start Filtering Data With the Help of Pandas Dataframe

Installing pandas

!pip install pandas

Importing the Pandas library, reading our sample data file and assigning it to “df” DataFrame

import pandas as pd
df = pd.read_csv(r"C:\Users\rajam\Desktop\sample_data.csv")

Let’s check out our dataframe:

print(df.head())

Sample_data

Now that we have our DataFrame, we will be applying various methods to filter it.

Method – 1: Filtering DataFrame by column value

We have a column named “Total_Sales” in our DataFrame and we want to filter out all the sales value which is greater than 300.

#Filter a DataFrame for a single column value with a given condition
 
greater_than = df[df['Total_Sales'] > 300]
print(greater_than.head())

Sales with Greater than 300

Method – 2: Filtering DataFrame based on multiple conditions

Here we are filtering all the values whose “Total_Sales” value is greater than 300 and also where the “Units” is greater than 20. We will have to use the python operator “&” which performs a bitwise AND operation in order to display the corresponding result.

#Filter a DataFrame with multiple conditions
 
filter_sales_units = df[(df['Total_Sales'] > 300) & (df["Units"] > 20)]
print(Filter_sales_units.head())

Filter on Sales and Units

Method – 3: Filtering DataFrame based on Date value

If we want to filter our data frame based on a certain date value, for example here we are trying to get all the results based on a particular date, in our case the results after the date ’03/10/21′.

#Filter a DataFrame based on specific date
 
date_filter = df[df['Date'] > '03/10/21']
print(date_filter.head())

Filter on Date

Method – 4: Filtering DataFrame based on Date value with multiple conditions

Here we are getting all the results for our Date operation evaluating multiple dates.

#Filter a DataFrame with multiple conditions our Date value
 
date_filter2 = df[(df['Date'] >= '3/25/2021') & (df['Date'] <'8/17/2021')]
print(date_filter2.head())

Filter on a date with multiple conditions

Method – 5: Filtering DataFrame based on a specific string

Here we are selecting a column called ‘Region’ and getting all the rows that are from the region ‘East’, thus filtering based on a specific string value.

#Filter a DataFrame to a specific string
 
east = df[df['Region'] == 'East']
print(east.head())

Filter based on a specific string

Method – 6: Filtering DataFrame based on a specific index value in a string

Here we are selecting a column called ‘Region’ and getting all the rows which has the letter ‘E’ as the first character i.e at index 0 in the specified column results.

#Filter a DataFrame to show rows starting with a specfic letter
 
starting_with_e = df[df['Region'].str[0]== 'E']
print(starting_with_e.head())

Filter based on a specific letter

Method – 7: Filtering DataFrame based on a list of values

Here we are filtering rows in the column ‘Region’ which contains the values ‘West’ as well as ‘East’ and display the combined result. Two methods can be used to perform this filtering namely using a pipe | operator with the corresponding desired set of values with the below syntax OR we can use the .isin() function to filter for the values in a given column, which in our case is the ‘Region’, and provide the list of the desired set of values inside it as a list.

#Filter a DataFrame rows based on list of values
 
#Method 1:
east_west = df[(df['Region'] == 'West') | (df['Region'] == 'East')]
print(east_west)
 
#Method 2:
east_west_1 = df[df['Region'].isin(['West', 'East'])]
print(east_west_1.head())

Output of Method -2

Method – 8: Filtering DataFrame rows based on specific values using RegEx

Here we want all the values in the column ‘Region’, which ends with ‘th’ in their string value and display them. In other words, we want our results to show the values of ‘North‘ and ‘South‘ and ignore ‘East’ and ‘West’. The method .str.contains() with the specified values along with the $ RegEx pattern can be used to get the desired results.

For more information please check the Regex Documentation

#Filtering the DataFrame rows using regular expressions(REGEX)
 
regex_df = df[df['Region'].str.contains('th$')]
print(regex_df.head())

Filter based on REGEX

Method – 9: Filtering DataFrame to check for null

Here, we’ll check for null and not null values in all the columns with the help of isnull() function.

#Filtering to check for null and not null values in all columns
 
df_null = df[df.isnull().any(axis=1)]
print(df_null.head())

Filter based on NULL or NOT null values

Method – 10: Filtering DataFrame to check for null values in a specific column.

#Filtering to check for null values if any in the 'Units' column
 
units_df = df[df['Units'].isnull()]
print(units_df.head())

Finding null values on specific columns

Method – 11: Filtering DataFrame to check for not null values in specific columns

#Filtering to check for not null values in the 'Units' column
 
df_not_null = df[df['Units'].notnull()]
print(df_not_null.head())

Finding not-null values on specific columns

Method – 12: Filtering DataFrame using query() with a condition

#Using query function in pandas
 
df_query = df.query('Total_Sales > 300')
print(df_query.head())

Filtering values with Query Function

Method – 13: Filtering DataFrame using query() with multiple conditions

#Using query function with multiple conditions in pandas
 
df_query_1 = df.query('Total_Sales > 300 and Units <18')
print(df_query_1.head())

Filtering multiple columns with Query Function

Method – 14: Filtering our DataFrame using the loc and iloc functions.

#Creating a sample DataFrame for illustrations
 
import numpy as np
data = pd.DataFrame({"col1" : np.arange(1, 20 ,2)}, index=[19, 18 ,8, 6, 0, 1, 2, 3, 4, 5])
print(data)

sample_data

Explanation: iloc considers rows based on the position of the given index, so that it takes only integers as values.

For more information please check out Pandas Documentation

#Filter with iloc
 
data.iloc[0 : 5]

Filter using iloc

Explanation: loc considers rows based on index labels

#Filter with loc
 
data.loc[0 : 5]

Filter using loc

You might be thinking about why the loc function returns 6 rows instead of 5 rows. This is because loc does not produce output based on index position. It considers labels of index only which can be an alphabet as well and includes both starting and endpoint.

Conclusion

So, these were some of the most common filtering methods used in pandas. There are many other filtering methods that could be used, but these are some of the most common.

Link: https://www.askpython.com/python-modules/pandas/filter-pandas-dataframe

#pandas #python #datafame

14 Cách Hàng đầu để Lọc Khung Dữ Liệu Pandas Một Cách Dễ Dàng

Bất cứ khi nào chúng tôi làm việc với bất kỳ loại dữ liệu nào, chúng tôi cần một bức tranh rõ ràng về loại dữ liệu mà chúng tôi đang xử lý. Đối với hầu hết dữ liệu ngoài kia, có thể chứa hàng nghìn hoặc thậm chí hàng triệu mục nhập với nhiều loại thông tin, thực sự không thể hiểu được dữ liệu đó nếu không có bất kỳ công cụ nào để trình bày dữ liệu ở định dạng ngắn gọn và dễ đọc.

Hầu hết thời gian chúng ta cần xem qua dữ liệu, thao tác và trực quan hóa nó để có được thông tin chi tiết. Chà, có một thư viện tuyệt vời mang tên gấu trúc cung cấp cho chúng ta khả năng đó. Thao tác thao tác dữ liệu thường xuyên nhất là Lọc dữ liệu. Nó rất giống với mệnh đề WHERE trong SQL hoặc bạn phải sử dụng một bộ lọc trong MS Excel để chọn các hàng cụ thể dựa trên một số điều kiện.

pandas là một công cụ phân tích / thao tác dữ liệu nguồn mở, linh hoạt và mạnh mẽ, về cơ bản là mộtgói pythoncung cấp tốc độ, tính linh hoạt và cấu trúc dữ liệu biểu cảm được tạo ra để làm việc với dữ liệu “quan hệ” hoặc “có nhãn” một cách trực quan và dễ dàng. Nó là một trong những thư viện phổ biến nhấtđể thực hiện phân tích dữ liệu trong thế giới thực bằng Python.

pandas được xây dựng dựa trên thư viện NumPy nhằm mục đích tích hợp tốt với môi trường máy tính khoa học và nhiều thư viện bên thứ 3 khác. Nó có hai cấu trúc dữ liệu chính là Series (1D) và Dataframe (2D) , trong hầu hết các trường hợp sử dụng trong thế giới thực là loại dữ liệu đang được xử lý trong nhiều lĩnh vực tài chính, máy tính khoa học, kỹ thuật và thống kê.

Hãy bắt đầu lọc dữ liệu với sự trợ giúp của khung dữ liệu Pandas

Cài đặt gấu trúc

!pip install pandas

Nhập thư viện Pandas, đọc tệp dữ liệu mẫu của chúng tôi và gán nó cho DataFrame “df”

import pandas as pd
df = pd.read_csv(r"C:\Users\rajam\Desktop\sample_data.csv")

Hãy kiểm tra khung dữ liệu của chúng tôi :

print(df.head())

Dữ liệu mẫu

Bây giờ chúng tôi đã có DataFrame của mình, chúng tôi sẽ áp dụng nhiều phương pháp khác nhau để lọc nó.

Phương pháp - 1 : Lọc DataFrame theo giá trị cột

Chúng tôi có một cột tên là “Total_Sales” trong DataFrame của mình và chúng tôi muốn lọc ra tất cả giá trị bán hàng lớn hơn 300.

#Filter a DataFrame for a single column value with a given condition
 
greater_than = df[df['Total_Sales'] > 300]
print(greater_than.head())

Doanh số lớn hơn 300

Phương pháp - 2 : Lọc DataFrame dựa trên nhiều điều kiện

Ở đây chúng tôi đang lọc tất cả các giá trị có giá trị “Total_Sales” lớn hơn 300 và cũng có giá trị “Đơn vị” lớn hơn 20. Chúng tôi sẽ phải sử dụng toán tử python “&” thực hiện thao tác AND bitwise để hiển thị kết quả tương ứng.

#Filter a DataFrame with multiple conditions
 
filter_sales_units = df[(df['Total_Sales'] > 300) & (df["Units"] > 20)]
print(Filter_sales_units.head())

Lọc theo Doanh số và Đơn vị

Phương pháp - 3 : Lọc DataFrame dựa trên giá trị Ngày tháng

Nếu chúng tôi muốn lọc khung dữ liệu của mình dựa trên một giá trị ngày nhất định, ví dụ: ở đây chúng tôi đang cố gắng lấy tất cả kết quả dựa trên một ngày cụ thể, trong trường hợp của chúng tôi là kết quả sau ngày '03/10/21'.

#Filter a DataFrame based on specific date
 
date_filter = df[df['Date'] > '03/10/21']
print(date_filter.head())

Lọc vào ngày

Phương pháp - 4: Lọc DataFrame dựa trên giá trị Ngày với nhiều điều kiện

Ở đây, chúng tôi nhận được tất cả các kết quả cho hoạt động Ngày đánh giá nhiều ngày của chúng tôi .

#Filter a DataFrame with multiple conditions our Date value
 
date_filter2 = df[(df['Date'] >= '3/25/2021') & (df['Date'] <'8/17/2021')]
print(date_filter2.head())

Lọc vào một ngày có nhiều điều kiện

Phương pháp - 5: Lọc DataFrame dựa trên một chuỗi cụ thể

Ở đây chúng tôi đang chọn một cột có tên là 'Khu vực' và lấy tất cả các hàng từ khu vực 'Đông', do đó lọc dựa trên một giá trị chuỗi cụ thể .

#Filter a DataFrame to a specific string
 
east = df[df['Region'] == 'East']
print(east.head())

Lọc dựa trên một chuỗi cụ thể

Phương pháp - 6: Lọc DataFrame dựa trên một giá trị chỉ mục cụ thể trong một chuỗi

Ở đây chúng tôi đang chọn một cột có tên là 'Vùng' và lấy tất cả các hàng có ký tự 'E' là ký tự đầu tiên, tức là ở chỉ số 0 trong kết quả cột được chỉ định.

#Filter a DataFrame to show rows starting with a specfic letter
 
starting_with_e = df[df['Region'].str[0]== 'E']
print(starting_with_e.head())

Lọc dựa trên một chữ cái cụ thể

Phương pháp - 7: Lọc DataFrame dựa trên danh sách các giá trị

Ở đây chúng tôi đang lọc các hàng trong cột 'Vùng' chứa các giá trị 'Tây' cũng như 'Đông' và hiển thị kết quả kết hợp. Hai phương pháp có thể được sử dụng để thực hiện việc lọc này là sử dụng đường ống | toán tử với tập giá trị mong muốn tương ứng với cú pháp bên dưới HOẶC chúng ta có thể sử dụng hàm .isin () để lọc các giá trị trong một cột nhất định, trong trường hợp của chúng ta là 'Vùng' và cung cấp danh sách tập hợp mong muốn của các giá trị bên trong nó dưới dạng danh sách.

#Filter a DataFrame rows based on list of values
 
#Method 1:
east_west = df[(df['Region'] == 'West') | (df['Region'] == 'East')]
print(east_west)
 
#Method 2:
east_west_1 = df[df['Region'].isin(['West', 'East'])]
print(east_west_1.head())

Đầu ra của Phương pháp -2

Phương pháp - 8: Lọc các hàng DataFrame dựa trên các giá trị cụ thể bằng cách sử dụng RegEx

Ở đây chúng tôi muốn tất cả các giá trị trong cột 'Vùng' , kết thúc bằng 'th' trong giá trị chuỗi của chúng và hiển thị chúng. Nói cách khác, chúng tôi muốn kết quả của mình hiển thị các giá trị của "Nor th " và "Sou th " và bỏ qua "East" và "West" . Phương thức .str.contains () với các giá trị được chỉ định cùng với mẫu $ RegEx có thể được sử dụng để có được kết quả mong muốn.

Để biết thêm thông tin, vui lòng kiểm tra Tài liệu Regex

#Filtering the DataFrame rows using regular expressions(REGEX)
 
regex_df = df[df['Region'].str.contains('th$')]
print(regex_df.head())

Lọc dựa trên REGEX

Phương pháp - 9: Lọc DataFrame để kiểm tra null

Ở đây, chúng tôi sẽ kiểm tra các giá trị null và không null trong tất cả các cột với sự trợ giúp của hàm isnull () .

#Filtering to check for null and not null values in all columns
 
df_null = df[df.isnull().any(axis=1)]
print(df_null.head())

Lọc dựa trên giá trị NULL hoặc NOT null

Phương pháp - 10: Lọc DataFrame để kiểm tra các giá trị null trong một cột cụ thể.

#Filtering to check for null values if any in the 'Units' column
 
units_df = df[df['Units'].isnull()]
print(units_df.head())

Tìm giá trị null trên các cột cụ thể

Phương pháp - 11: Lọc DataFrame để kiểm tra các giá trị không rỗng trong các cột cụ thể

#Filtering to check for not null values in the 'Units' column
 
df_not_null = df[df['Units'].notnull()]
print(df_not_null.head())

Tìm các giá trị not-null trên các cột cụ thể

Phương pháp - 12: Lọc DataFrame bằng cách sử dụng query()với một điều kiện

#Using query function in pandas
 
df_query = df.query('Total_Sales > 300')
print(df_query.head())

Lọc các giá trị bằng QueryHàm

Phương pháp - 13: Lọc DataFrame bằng query()nhiều điều kiện

#Using query function with multiple conditions in pandas
 
df_query_1 = df.query('Total_Sales > 300 and Units <18')
print(df_query_1.head())

Lọc nhiều cột với QueryHàm

Phương pháp - 14: Lọc DataFrame của chúng tôi bằng cách sử dụng các hàm locvà iloc.

#Creating a sample DataFrame for illustrations
 
import numpy as np
data = pd.DataFrame({"col1" : np.arange(1, 20 ,2)}, index=[19, 18 ,8, 6, 0, 1, 2, 3, 4, 5])
print(data)

dữ liệu mẫu

Giải thích : iloc xem xét các hàng dựa trên vị trí của chỉ mục đã cho, do đó nó chỉ nhận các số nguyên làm giá trị.

Để biết thêm thông tin, vui lòng xem Tài liệu về Gấu trúc

#Filter with iloc
 
data.iloc[0 : 5]

Lọc bằng cách sử dụngiloc

Giải thích : loc xem xét các hàng dựa trên nhãn chỉ mục

#Filter with loc
 
data.loc[0 : 5]

Lọc bằng cách sử dụngloc

Bạn có thể đang suy nghĩ về lý do tại sao lochàm trả về 6 hàng thay vì 5 hàng. Điều này là do không tạo ra sản lượng dựa trên vị trí chỉ mục. Nó chỉ xem xét các nhãn của chỉ mục cũng có thể là một bảng chữ cái và bao gồm cả điểm đầu và điểm cuối. loc 

Sự kết luận

Vì vậy, đây là một số phương pháp lọc phổ biến nhất được sử dụng ở gấu trúc. Có nhiều phương pháp lọc khác có thể được sử dụng, nhưng đây là một số phương pháp phổ biến nhất.

Liên kết: https://www.askpython.com/python-modules/pandas/filter-pandas-dataframe

#pandas #python #datafame

14 лучших способов легко фильтровать кадры данных Pandas

Всякий раз, когда мы работаем с данными любого рода, нам нужна четкая картина того, с какими данными мы имеем дело. Для большинства имеющихся данных, которые могут содержать тысячи или даже миллионы записей с разнообразной информацией, действительно невозможно разобраться в этих данных без какого-либо инструмента для представления данных в кратком и удобочитаемом формате.

Большую часть времени нам нужно просматривать данные, манипулировать ими и визуализировать их для получения информации. Что ж, есть отличная библиотека под названием pandas, которая предоставляет нам эту возможность. Наиболее частой операцией манипулирования данными является фильтрация данных. Это очень похоже на предложение WHERE в SQL, или вы должны были использовать фильтр в MS Excel для выбора определенных строк на основе некоторых условий.

pandas — это мощный, гибкий инструмент с открытым исходным кодом для анализа/манипулирования данными, который, по сути, представляет собойпакет Python, обеспечивающий скорость, гибкость и выразительные структуры данных, созданные для интуитивно понятной и простой работы с «реляционными» или «помеченными» данными. Это одна из самых популярных библиотекдля реального анализа данных в Python.

pandas построен на основе библиотеки NumPy, которая нацелена на хорошую интеграцию с научной вычислительной средой и множеством других сторонних библиотек. Он имеет две основные структуры данных, а именно Series (1D) и Dataframes(2D) , которые в большинстве реальных случаев использования представляют собой тип данных, с которыми имеют дело во многих секторах финансов, научных вычислений, инженерии и статистики.

Давайте начнем фильтровать данные с помощью Pandas Dataframe

Установка панд

!pip install pandas

Импорт библиотеки Pandas, чтение нашего примера файла данных и назначение его в «df» DataFrame

import pandas as pd
df = pd.read_csv(r"C:\Users\rajam\Desktop\sample_data.csv")

Давайте проверим наш фрейм данных :

print(df.head())

Образец данных

Теперь, когда у нас есть DataFrame, мы будем применять различные методы для его фильтрации.

Метод — 1 : фильтрация DataFrame по значению столбца

У нас есть столбец с именем «Total_Sales» в нашем DataFrame, и мы хотим отфильтровать все значения продаж, превышающие 300.

#Filter a DataFrame for a single column value with a given condition
 
greater_than = df[df['Total_Sales'] > 300]
print(greater_than.head())

Продажи с более чем 300

Метод — 2 : фильтрация DataFrame на основе нескольких условий

Здесь мы фильтруем все значения, у которых значение «Total_Sales» больше 300, а также где «Единицы» больше 20. Нам нужно будет использовать оператор Python «&», который выполняет побитовую операцию И, чтобы отобразить соответствующий результат.

#Filter a DataFrame with multiple conditions
 
filter_sales_units = df[(df['Total_Sales'] > 300) & (df["Units"] > 20)]
print(Filter_sales_units.head())

Фильтр по продажам и единицам

Метод — 3 : фильтрация DataFrame на основе значения даты

Если мы хотим отфильтровать наш фрейм данных на основе определенного значения даты, например, здесь мы пытаемся получить все результаты на основе определенной даты, в нашем случае результаты после даты «10.03.21».

#Filter a DataFrame based on specific date
 
date_filter = df[df['Date'] > '03/10/21']
print(date_filter.head())

Фильтр по дате

Метод — 4: фильтрация DataFrame на основе значения даты с несколькими условиями

Здесь мы получаем все результаты нашей операции Date, оценивающей несколько дат .

#Filter a DataFrame with multiple conditions our Date value
 
date_filter2 = df[(df['Date'] >= '3/25/2021') & (df['Date'] <'8/17/2021')]
print(date_filter2.head())

Фильтр по дате с несколькими условиями

Метод — 5: фильтрация DataFrame на основе определенной строки

Здесь мы выбираем столбец под названием «Регион» и получаем все строки из региона «Восток», таким образом фильтруя на основе определенного строкового значения .

#Filter a DataFrame to a specific string
 
east = df[df['Region'] == 'East']
print(east.head())

Фильтровать по определенной строке

Метод — 6: фильтрация DataFrame на основе определенного значения индекса в строке

Здесь мы выбираем столбец под названием «Регион» и получаем все строки, в которых буква «Е» является первым символом, т.е. индексом 0 в результатах указанного столбца.

#Filter a DataFrame to show rows starting with a specfic letter
 
starting_with_e = df[df['Region'].str[0]== 'E']
print(starting_with_e.head())

Фильтр по определенной букве

Метод — 7: Фильтрация DataFrame на основе списка значений

Здесь мы фильтруем строки в столбце «Регион», который содержит значения «Запад», а также «Восток», и отображаем объединенный результат. Для выполнения этой фильтрации можно использовать два метода, а именно использование канала | оператор с соответствующим желаемым набором значений с приведенным ниже синтаксисом ИЛИ мы можем использовать функцию .isin() для фильтрации значений в данном столбце, которым в нашем случае является «Регион», и предоставить список желаемого набора значений внутри него в виде списка.

#Filter a DataFrame rows based on list of values
 
#Method 1:
east_west = df[(df['Region'] == 'West') | (df['Region'] == 'East')]
print(east_west)
 
#Method 2:
east_west_1 = df[df['Region'].isin(['West', 'East'])]
print(east_west_1.head())

Выход метода -2

Метод — 8: фильтрация строк DataFrame на основе определенных значений с использованием RegEx

Здесь нам нужны все значения в столбце «Регион» , которые заканчиваются на «th» в их строковом значении, и отобразить их. Другими словами, мы хотим, чтобы наши результаты показывали значения «Север » и «Юг » и игнорировали «Восток» и «Запад» . Метод .str.contains() с указанными значениями вместе с шаблоном $ RegEx можно использовать для получения желаемых результатов.

Для получения дополнительной информации ознакомьтесь с документацией по регулярным выражениям.

#Filtering the DataFrame rows using regular expressions(REGEX)
 
regex_df = df[df['Region'].str.contains('th$')]
print(regex_df.head())

Фильтр на основе REGEX

Метод — 9: фильтрация DataFrame для проверки на нуль

Здесь мы проверим нулевые и не нулевые значения во всех столбцах с помощью функции isnull() .

#Filtering to check for null and not null values in all columns
 
df_null = df[df.isnull().any(axis=1)]
print(df_null.head())

Фильтр на основе значений NULL или NOT null

Метод — 10: фильтрация DataFrame для проверки нулевых значений в определенном столбце.

#Filtering to check for null values if any in the 'Units' column
 
units_df = df[df['Units'].isnull()]
print(units_df.head())

Поиск нулевых значений в определенных столбцах

Метод — 11: фильтрация DataFrame для проверки ненулевых значений в определенных столбцах

#Filtering to check for not null values in the 'Units' column
 
df_not_null = df[df['Units'].notnull()]
print(df_not_null.head())

Поиск ненулевых значений в определенных столбцах

Метод — 12: Фильтрация DataFrame query()с использованием условия

#Using query function in pandas
 
df_query = df.query('Total_Sales > 300')
print(df_query.head())

Фильтрация значений с Queryфункцией

Метод — 13: фильтрация DataFrame с использованием query()нескольких условий

#Using query function with multiple conditions in pandas
 
df_query_1 = df.query('Total_Sales > 300 and Units <18')
print(df_query_1.head())

Фильтрация нескольких столбцов с Queryфункцией

Метод — 14: фильтрация нашего DataFrame с использованием функций locи .iloc

#Creating a sample DataFrame for illustrations
 
import numpy as np
data = pd.DataFrame({"col1" : np.arange(1, 20 ,2)}, index=[19, 18 ,8, 6, 0, 1, 2, 3, 4, 5])
print(data)

образец данных

Объяснение : iloc считает строки на основе позиции заданного индекса, поэтому в качестве значений принимает только целые числа.

Для получения дополнительной информации ознакомьтесь с документацией Pandas.

#Filter with iloc
 
data.iloc[0 : 5]

Фильтровать с помощьюiloc

Объяснение : loc считает строки на основе меток индекса .

#Filter with loc
 
data.loc[0 : 5]

Фильтровать с помощьюloc

Вы можете подумать, почему locфункция возвращает 6 строк вместо 5 строк. Это связано с тем , что вывод не производится на основе позиции индекса. Он рассматривает только метки индекса, которые также могут быть алфавитом, и включает как начальную, так и конечную точку. loc 

Вывод

Итак, это были одни из наиболее распространенных методов фильтрации, используемых в пандах. Существует множество других методов фильтрации, которые можно использовать, но эти являются одними из наиболее распространенных.

Ссылка: https://www.askpython.com/python-modules/pandas/filter-pandas-dataframe

#pandas #python #datafame

14 Meilleures Façons De Filtrer Facilement Les Dataframes Pandas

Chaque fois que nous travaillons avec des données de toutes sortes, nous avons besoin d'une image claire du type de données avec lesquelles nous traitons. Pour la plupart des données disponibles, qui peuvent contenir des milliers, voire des millions d'entrées avec une grande variété d'informations, il est vraiment impossible de donner un sens à ces données sans aucun outil pour présenter les données dans un format court et lisible.

La plupart du temps, nous devons parcourir les données, les manipuler et les visualiser pour obtenir des informations. Eh bien, il existe une excellente bibliothèque qui porte le nom de pandas et qui nous offre cette capacité. L'opération de manipulation de données la plus fréquente est le filtrage de données. Il est très similaire à la clause WHERE dans SQL ou vous devez avoir utilisé un filtre dans MS Excel pour sélectionner des lignes spécifiques en fonction de certaines conditions.

pandas est un outil d'analyse/manipulation de données puissant, flexible et open source qui est essentiellement unpackage pythonqui offre vitesse, flexibilité et structures de données expressives conçues pour fonctionner avec des données « relationnelles » ou « étiquetées » de manière intuitive et simple. C'est l'une des bibliothèques les plus populairespour effectuer une analyse de données du monde réel en Python.

pandas est construit au-dessus de la bibliothèque NumPy qui vise à bien s'intégrer à l'environnement informatique scientifique et à de nombreuses autres bibliothèques tierces. Il comporte deux structures de données principales, à savoir Series (1D) et Dataframes (2D) , qui, dans la plupart des cas d'utilisation réels, correspondent au type de données traitées dans de nombreux secteurs de la finance, du calcul scientifique, de l'ingénierie et des statistiques.

Commençons à filtrer les données à l'aide de Pandas Dataframe

Installer des pandas

!pip install pandas

Importation de la bibliothèque Pandas, lecture de notre exemple de fichier de données et affectation à "df" DataFrame

import pandas as pd
df = pd.read_csv(r"C:\Users\rajam\Desktop\sample_data.csv")

Voyons notre dataframe :

print(df.head())

Sample_data

Maintenant que nous avons notre DataFrame, nous allons appliquer différentes méthodes pour le filtrer.

Méthode – 1 : Filtrage de DataFrame par valeur de colonne

Nous avons une colonne nommée "Total_Sales" dans notre DataFrame et nous voulons filtrer toute la valeur des ventes supérieure à 300.

#Filter a DataFrame for a single column value with a given condition
 
greater_than = df[df['Total_Sales'] > 300]
print(greater_than.head())

Ventes avec plus de 300

Méthode – 2 : Filtrage de DataFrame basé sur plusieurs conditions

Ici, nous filtrons toutes les valeurs dont la valeur "Total_Sales" est supérieure à 300 et également où les "Unités" sont supérieures à 20. Nous devrons utiliser l'opérateur python "&" qui effectue une opération ET au niveau du bit afin d'afficher le résultat correspondant.

#Filter a DataFrame with multiple conditions
 
filter_sales_units = df[(df['Total_Sales'] > 300) & (df["Units"] > 20)]
print(Filter_sales_units.head())

Filtrer sur les ventes et les unités

Méthode - 3 : Filtrage de DataFrame basé sur la valeur Date

Si nous voulons filtrer notre trame de données en fonction d'une certaine valeur de date, par exemple ici nous essayons d'obtenir tous les résultats en fonction d'une date particulière, dans notre cas les résultats après la date '03/10/21'.

#Filter a DataFrame based on specific date
 
date_filter = df[df['Date'] > '03/10/21']
print(date_filter.head())

Filtrer par date

Méthode - 4 : Filtrage de DataFrame en fonction de la valeur Date avec plusieurs conditions

Ici, nous obtenons tous les résultats de notre opération Date évaluant plusieurs dates .

#Filter a DataFrame with multiple conditions our Date value
 
date_filter2 = df[(df['Date'] >= '3/25/2021') & (df['Date'] <'8/17/2021')]
print(date_filter2.head())

Filtrer sur une date avec plusieurs conditions

Méthode - 5 : Filtrage de DataFrame en fonction d'une chaîne spécifique

Ici, nous sélectionnons une colonne appelée 'Region' et obtenons toutes les lignes qui proviennent de la région 'East', filtrant ainsi en fonction d'une valeur de chaîne spécifique .

#Filter a DataFrame to a specific string
 
east = df[df['Region'] == 'East']
print(east.head())

Filtre basé sur une chaîne spécifique

Méthode - 6 : Filtrage de DataFrame en fonction d'une valeur d'index spécifique dans une chaîne

Ici, nous sélectionnons une colonne appelée 'Region' et obtenons toutes les lignes qui ont la lettre 'E' comme premier caractère, c'est-à-dire à l'index 0 dans les résultats de colonne spécifiés.

#Filter a DataFrame to show rows starting with a specfic letter
 
starting_with_e = df[df['Region'].str[0]== 'E']
print(starting_with_e.head())

Filtre basé sur une lettre spécifique

Méthode - 7 : Filtrage de DataFrame basé sur une liste de valeurs

Ici, nous filtrons les lignes dans la colonne « Région » qui contient les valeurs « Ouest » ainsi que « Est » et affichons le résultat combiné. Deux méthodes peuvent être utilisées pour effectuer ce filtrage à savoir l'utilisation d'un tube | opérateur avec l'ensemble de valeurs souhaité correspondant avec la syntaxe ci-dessous OU nous pouvons utiliser la fonction .isin() pour filtrer les valeurs dans une colonne donnée, qui dans notre cas est la 'Région', et fournir la liste de l'ensemble souhaité de valeurs à l'intérieur sous forme de liste.

#Filter a DataFrame rows based on list of values
 
#Method 1:
east_west = df[(df['Region'] == 'West') | (df['Region'] == 'East')]
print(east_west)
 
#Method 2:
east_west_1 = df[df['Region'].isin(['West', 'East'])]
print(east_west_1.head())

Sortie de la méthode -2

Méthode - 8: Filtrage des lignes DataFrame en fonction de valeurs spécifiques à l'aide de RegEx

Ici, nous voulons toutes les valeurs de la colonne 'Region' , qui se termine par 'th' dans leur valeur de chaîne et les afficher. En d'autres termes, nous voulons que nos résultats montrent les valeurs de « Nord » et « Sud » et ignorent « Est » et « Ouest » . La méthode .str.contains() avec les valeurs spécifiées avec le modèle $ RegEx peut être utilisée pour obtenir les résultats souhaités.

Pour plus d'informations, veuillez consulter la documentation Regex

#Filtering the DataFrame rows using regular expressions(REGEX)
 
regex_df = df[df['Region'].str.contains('th$')]
print(regex_df.head())

Filtre basé sur REGEX

Méthode - 9: Filtrage de DataFrame pour vérifier null

Ici, nous allons vérifier les valeurs nulles et non nulles dans toutes les colonnes à l'aide de la fonction isnull() .

#Filtering to check for null and not null values in all columns
 
df_null = df[df.isnull().any(axis=1)]
print(df_null.head())

Filtre basé sur les valeurs NULL ou NOT null

Méthode - 10 : Filtrage de DataFrame pour vérifier les valeurs nulles dans une colonne spécifique.

#Filtering to check for null values if any in the 'Units' column
 
units_df = df[df['Units'].isnull()]
print(units_df.head())

Recherche de valeurs nulles sur des colonnes spécifiques

Méthode - 11 : Filtrage de DataFrame pour vérifier les valeurs non nulles dans des colonnes spécifiques

#Filtering to check for not null values in the 'Units' column
 
df_not_null = df[df['Units'].notnull()]
print(df_not_null.head())

Recherche de valeurs non nulles sur des colonnes spécifiques

Méthode - 12: Filtrage de DataFrame à l'aide query()d'une condition

#Using query function in pandas
 
df_query = df.query('Total_Sales > 300')
print(df_query.head())

Filtrer les valeurs avec Queryla fonction

Méthode - 13: Filtrage de DataFrame à l'aide query()de plusieurs conditions

#Using query function with multiple conditions in pandas
 
df_query_1 = df.query('Total_Sales > 300 and Units <18')
print(df_query_1.head())

Filtrer plusieurs colonnes avec QueryFunction

Méthode – 14 : Filtrage de notre DataFrame à l'aide des fonctions locet iloc.

#Creating a sample DataFrame for illustrations
 
import numpy as np
data = pd.DataFrame({"col1" : np.arange(1, 20 ,2)}, index=[19, 18 ,8, 6, 0, 1, 2, 3, 4, 5])
print(data)

sample_data

Explication : iloc considère les lignes en fonction de la position de l'index donné, de sorte qu'il ne prend que des entiers comme valeurs.

Pour plus d'informations, veuillez consulter la documentation de Pandas

#Filter with iloc
 
data.iloc[0 : 5]

Filtrer en utilisantiloc

Explication : loc considère les lignes en fonction des étiquettes d'index

#Filter with loc
 
data.loc[0 : 5]

Filtrer en utilisantloc

Vous vous demandez peut-être pourquoi la locfonction renvoie 6 lignes au lieu de 5 lignes. En effet , ne produit pas de sortie basée sur la position de l'index. Il ne prend en compte que les étiquettes d'index qui peuvent également être un alphabet et incluent à la fois le point de départ et le point final. loc 

Conclusion

Donc, ce sont quelques-unes des méthodes de filtrage les plus couramment utilisées dans les pandas. Il existe de nombreuses autres méthodes de filtrage qui pourraient être utilisées, mais celles-ci sont parmi les plus courantes.

Lien : https://www.askpython.com/python-modules/pandas/filter-pandas-dataframe

#pandas #python #datafame