Reorder columns in returned validated dataframe

[JeremyL-01]

My team heavily use Pandera to validate our data coming into our transformations as well as "rubber-stamp" the dataframes we generate as output.

Our output dataframes have to explicitly adhere to the datatypes AND column order of our defined Pandera schema.

We use the strict="filter" to ensure we cut out any columns not specified in our schema. We use coerce=True to ensure our validated dataframe has the correct data types. However, there's nothing built in (that we've found) to reorder the dataframe columns to align with the schema. Any reason this can't be done automatically as part of the returned validated dataframe? Or perhaps an option could be added?

We currently use this manual work-around every time to fix the column order (which, coincidently, negates the usefulness of strict="filter") :
df = df[schema.columns.keys()]

Example:

import pandas as pd
import pandera as pa

schema = pa.DataFrameSchema(
    {
        "column2": pa.Column(str), 
        "column1": pa.Column(int)
     }, 
     strict="filter",
     coerce=True
)

df = pd.DataFrame(
    {
        "column1": [1.0, 2.0, 3.0], 
        "column2": ["a", "b", "c"], 
        "column3": [1.1, 2.2, 3.3]
     }
)

validated_df = schema.validate(df)
validated_df.info()

Validated dataframe doesn't have columns in order specified in schema:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3 entries, 0 to 2
Data columns (total 2 columns):
 #   Column   Non-Null Count  Dtype 
---  ------   --------------  ----- 
 0   column1  3 non-null      int64 
 1   column2  3 non-null      object
dtypes: int64(1), object(1)
memory usage: 176.0+ bytes

Manual work-around:

validated_df = validated_df[schema.columns.keys()]

validated_df.info()

Validated dataframe now has columns in the order specified in the schema:

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 3 entries, 0 to 2
Data columns (total 2 columns):
 #   Column   Non-Null Count  Dtype 
---  ------   --------------  ----- 
 0   column2  3 non-null      object
 1   column1  3 non-null      int64 
dtypes: int64(1), object(1)
memory usage: 176.0+ bytes

[JeremyL-01]

Duplicate conversation occurring here: #1317