81 lines
2.6 KiB
Python
81 lines
2.6 KiB
Python
from typing import Any, Dict
|
|
import pandas as pd
|
|
from .utils import column_labels_string_to_list
|
|
|
|
|
|
class PandasValueCounts:
|
|
"""
|
|
Creates a frequency table from one or more columns in a DataFrame.
|
|
|
|
|
|
|
|
category: Categorical summary
|
|
"""
|
|
|
|
@classmethod
|
|
def INPUT_TYPES(cls) -> Dict[str, Any]:
|
|
"""
|
|
Defines the input types for the function.
|
|
|
|
Returns:
|
|
Dict[str, Any]: A dictionary specifying required input types.
|
|
"""
|
|
return {
|
|
"required": {
|
|
"dataframe": ("DATAFRAME", {}),
|
|
"column_names": ("STRING", {"default": ""})
|
|
}
|
|
}
|
|
|
|
RETURN_TYPES: tuple = ("DATAFRAME",)
|
|
FUNCTION: str = "f"
|
|
CATEGORY: str = "Data Analysis"
|
|
|
|
def f(self, dataframe: pd.DataFrame, column_names: str) -> tuple:
|
|
"""
|
|
Creates a frequency tableConverts one or more columns in a DataFrame.
|
|
|
|
Args:
|
|
dataframe (DataFrame): The DataFrame.
|
|
column_names (str): A comma-separated list of column names containing categorical data.
|
|
|
|
Returns:
|
|
tuple: A tuple containing the DataFrame with selected columns.
|
|
"""
|
|
|
|
# Note
|
|
# If the original DataFrame contains a single column with categorical variables:
|
|
#
|
|
# Category
|
|
# --------
|
|
# A
|
|
# A
|
|
# A
|
|
# B
|
|
# B
|
|
# C
|
|
# C
|
|
#
|
|
# and you want to compute the count of each variable and call value_counts().
|
|
# After value_counts(), the column entries are became the index in a generated Series:
|
|
# df["Category"].value_counts().to_frame()
|
|
# Category
|
|
# A 3
|
|
# B 2
|
|
# C 2
|
|
#
|
|
# After you call reset_index, A, B, C are converted back to a column, but the column label
|
|
# is set to "index".
|
|
# Therefore, you need to do the following:
|
|
# * Convert a Series to a DataFrame (to_frame)
|
|
# * Rename the count column to Count (rename)
|
|
# * Convert the original column from index to a column (reset_index)
|
|
# * Restore the original column label (rename)
|
|
cols = column_labels_string_to_list(dataframe, column_names)
|
|
if len(cols) == 1:
|
|
col_label = cols[0]
|
|
df2 = dataframe[col_label].value_counts().to_frame().rename(columns={col_label:"Count"}).reset_index().rename(columns={"index":col_label})
|
|
else:
|
|
df2 = dataframe[cols].value_counts().to_frame().rename(columns={0:"Count"}).reset_index()
|
|
return (df2,)
|