pywrangle 0.2.52

Auxiliary functions to clean pandas data frames

Pywrangle

Library for Python data wrangling to streamline string cleaning, identifying missing data, and tracking dataframe changes. Available on PyPI here

Install

• Python 3.6+
• numpy
• pandas

To install pywrangle, use pip:

pip install pywrangle

Import

Per convention with Python Analysis modules, import pywrangle as follows:

>>> import pywrangle as pw

String cleaning

def clean_str_columns(df: object, col_strcase_tuple: tuple) -> df:

Master function to clean string columns using col_strcase_tuple key.

col_strcase_tuple is a tuple of tuples representing the column names to be cleaned and an ordinal number for the pandas str cleaning method to use. Ordinal case control structure to determine case: 0 : lower_case 1 : title_case 2 : upper_case

df_winereviews = pd.read_csv("../input/wine-reviews/winemag-data_first150k.csv")
col_strcase_tuple = (
        ("country", 2),
        ("description", 0)
        ("province", 1),
    )
df_winereviews = clean_str_columns( df_winereviews, col_strcase_tuple)
    column name:		str.clean_method
    country     		upper
    description 		lower
    province    		title

Missing Data

print_nulls_per_col(df) -> None:

Calculates number of null values in each column and prints result.

Dataframe changes

The dataframe change functions record_df_info and print_df_changes are used in conjunction.

>>> old_df = pw.df_info(df)
>>> ... # some change to df
>>> pw.print_df_changes(df, old_df)

record_df_info(df, _name: str = "before") -> None:

Records information about the dataframe.

Information includes: - name (state of the dict, before or after) - number of rows - number of columns - size of df

recorded dataframe information is passed to compare_dfs() to check differences between dataframes.

print_df_changes( df, dict_recorded_info: dict, show_col_names: bool = False ) -> None:

Prints differences between dataframe and previously recorded information.

History

version = "0.2.40"

• refactored code for clarity
• added display info to print_df_changes

version = "0.2.1"

• Created init file for function imports
• Documentation on importing pywrangle
• Added numpy as required package.
• Changed package requirements to greater than or equal to.

version = "0.0.1"

• Init