pyspark-json-loader 0.1.5

A package to load and preprocess JSON data using PySpark

PySpark JSON Loader

Overview

pyspark-json-loader is a Python package designed to facilitate loading and preprocessing JSON data using PySpark. It provides functions to start a Spark session, connect to a PostgreSQL database, preprocess data, and convert Spark DataFrames to Pandas DataFrames.

Installation

To install the package, run:

pip install pyspark-json-loader

Usage

Here is a detailed guide on how to use the functions provided by the pyspark-json-loader package.

Importing the Module

To use the functions in this package, you need to import them as follows:

from pyspark_json_loader import load_json_file, start_spark_session, connect_to_db, preprocess_dataframe, convert_to_pandas

Functions

1. load_json_file(file_name)

Description: This function loads a JSON file and returns its contents as a Python dictionary.

Parameters:

• file_name (str): The path to the JSON file.

Returns:

• dict: The contents of the JSON file.

Example:

json_data = load_json_file('data.json')
print(json_data)

Output:

{
    "column1": "value1",
    "column2": "value2"
}

2. start_spark_session()

Description: This function starts a Spark session with the specified JAR file.

Returns:

• SparkSession: The Spark session object.

Example:

spark = start_spark_session()
print(spark)

Output:

<pyspark.sql.session.SparkSession object at 0x...>

3. connect_to_db(spark, host_name, port_number, db_name, user, password, query, null_value=0)

Description: This function connects to a PostgreSQL database using the provided connection details and query.

Parameters:

• spark (SparkSession): The Spark session object.
• host_name (str): The hostname of the PostgreSQL server.
• port_number (int): The port number of the PostgreSQL server.
• db_name (str): The name of the database.
• user (str): The username for the database.
• password (str): The password for the database.
• query (str): The SQL query to execute.
• null_value (int, optional): The value to use for null values. Default is 0.

Returns:

• DataFrame: The resulting DataFrame from the query.

Example:

df = connect_to_db(spark, 'localhost', 5432, 'mydatabase', 'user', 'password', 'SELECT * FROM mytable')
df.show()

Output:

| id | name |
|----|------|
| 1  | John |
| 2  | Jane |

4. preprocess_dataframe(df, json_data)

Description: This function preprocesses a DataFrame by filling null values based on the provided JSON data.

Parameters:

• df (DataFrame): The Spark DataFrame to preprocess.
• json_data (dict): The JSON data to use for preprocessing.

Returns:

• DataFrame: The preprocessed DataFrame.

Example:

preprocessed_df = preprocess_dataframe(df, json_data)
preprocessed_df.show()

Output:

+---+------+
| id|  name|
+---+------+
|  1|  John|
|  2|  Jane|
+---+------+

5. convert_to_pandas(grouped_metrics_df)

Description: This function converts a Spark DataFrame to a Pandas DataFrame and adds a UUID column.

Parameters:

• grouped_metrics_df (DataFrame): The Spark DataFrame to convert.

Returns:

• DataFrame: The resulting Pandas DataFrame.

Example:

pandas_df = convert_to_pandas(preprocessed_df)
print(pandas_df)

Output:

   id   name                               Id
0   1   John  0a539f3c... (UUID)
1   2   Jane  1d2e4f5a... (UUID)