modin.pandas.DataFrame.cache_result¶

DataFrame.cache_result(inplace: bool = False) → Optional[DataFrame][source]¶

Persists the current Snowpark pandas DataFrame to a temporary table to improve the latency of subsequent operations.

Parameters:

inplace – bool, default False Whether to perform the materialization inplace.

Returns:

Snowpark pandas DataFrame or None

Cached Snowpark pandas DataFrame or None if inplace=True.

Note

• The temporary table produced by this method lasts for the duration of the session.

Examples:

Let’s make a DataFrame using a computationally expensive operation, e.g.: >>> df = pd.concat([pd.DataFrame([range(i, i+5)]) for i in range(0, 150, 5)])

Due to Snowpark pandas lazy evaluation paradigm, every time this DataFrame is used, it will be recomputed - causing every subsequent operation on this DataFrame to re-perform the 30 unions required to produce it. This makes subsequent operations more expensive. The cache_result API can be used to persist the DataFrame to a temporary table for the duration of the session - replacing the nested 30 unions with a single read from a table.

>>> new_df = df.cache_result()

Copy

>>> import numpy as np

Copy

>>> np.all((new_df == df).values)
True

Copy

>>> df.reset_index(drop=True, inplace=True) # Slower

Copy

>>> new_df.reset_index(drop=True, inplace=True) # Faster

Copy