dotools_py.get.subset_df

Contents

dotools_py.get.subset_df#

dotools_py.get.subset_df(df, col=None, col_groups=None, comparison='include')[source]#

Subset Pandas DataFrame.

Subset a Pandas DataFrame based on one or multiple columns. To subset for multiple columns, provide a list to col and col_groups in the same order. For example, you have a dataframe df with two columns: sex that conteins “male” and “female” and age with values from 20 to 90. You can subset to select only males below 50 providing: col = ["sex", "age"] and col_groups = ["male", 50] and then specifying the type of comparison for each column: comparison = ["include", "<"]. More examples in the section below.

Parameters:
df DataFrame

Pandas DataFrame

col str | list | None (default: None)

Name of the columns to use for subseting. To subset based on multiple columns provide a list.

col_groups str | list | float | bool | None (default: None)

Values to use for subsetting each column. If several columns are provided, provide a list.

comparison Union[Literal['>=', '>', '==', '<', '<=', 'include', 'exclude'], list] (default: 'include')

How the subset will be performed. If several columns are provided, provide a list.

Return type:

DataFrame

Returns:

Returns a pandas DataFrame subsetted.

Examples

>>> import dotools_py as do
>>> adata = do.dt.example_10x_processed()
>>> df = adata.obs.copy()
>>> df.shape
(700, 22)
>>> df_subset = subset_df(df, col=["condition", "annotation"],  col_groups=["healthy", "NK"], comparison="include")
>>> df_subset.shape
(111, 22)
>>> df_subset.value_counts(["condition", "annotation"])
condition  annotation
healthy    NK            111
Name: count, dtype: int64
>>> df_subset = subset_df(df, col=["total_counts", "annotation"], col_groups=[1000, ["B_cells", "T_cells"]], comparison=["<", "include"] )
>>> df_subset.shape
(1, 22)
>>> df_subset.head(5)[["total_counts", "annotation"]].sort_values("total_counts", ascending=False)
                               total_counts annotation
GGAGAACCAAAGCTCT-1-batch2         938.0    T_cells
Contents