(TIL) Pandas: Use dtype to speed up reading with read_csv

less than 1 minute read

By default, pandas will infer the data types of the columns when reading in a csv file. To speed up this read, you can specify the data types using the dtype argument of read_csv().

For example:

df = pd \
    .read_csv(
        file.csv,
        dtype={
            'a': int,
            'b': np.float64,
            'c': str,
            },
        )

Tags: ,

Categories:

Updated:

Comments