Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

PERF: read_csv() very slow after updating to version 1.3.4 for CSV file with large number of columns #44106

Closed
2 of 3 tasks
haghshenas opened this issue Oct 19, 2021 · 3 comments · Fixed by #44192
Closed
2 of 3 tasks

PERF: read_csv() very slow after updating to version 1.3.4 for CSV file with large number of columns #44106

haghshenas opened this issue Oct 19, 2021 · 3 comments · Fixed by #44192
Labels
Needs Triage Issue that has not been reviewed by a pandas team member Performance Memory or execution speed performance
Milestone
1.3.5

Comments

@haghshenas
Copy link

  • I have checked that this issue has not already been reported.

  • I have confirmed this issue exists on the latest version of pandas.

  • I have confirmed this issue exists on the master branch of pandas.

Reproducible Example

For the sake of this bug report, I created a relatively small CSV file of random numbers. My real CSV is of shape (200, 662688) and for that, updating to pandas 1.3.4 increased the time for loading from 1 minute to over 2 hours!
My investigation showed all 1.3.* versions have the same issue.

import os
import pandas
import numpy
import timeit

def generate_sample():
    if os.path.exists("test_small.csv.gz") == False:
        nb_col = 100000
        nb_row = 5
        feature_list = {'sample': ['s_' + str(i+1) for i in range(nb_row)]}
        for i in range(nb_col):
            feature_list.update({'feature_' + str(i+1): list(numpy.random.uniform(low=0, high=10, size=nb_row))})
        df = pandas.DataFrame(feature_list)
        df.to_csv("test_small.csv.gz", index=False, float_format="%.6f")

def load_csv_file():
    col_names = pandas.read_csv("test_small.csv.gz", low_memory=False, nrows=1).columns
    types_dict = {col: numpy.float32 for col in col_names}
    types_dict.update({'sample': str})
    feature_df = pandas.read_csv("test_small.csv.gz", index_col="sample", na_filter=False, dtype=types_dict, low_memory=False)
    print("loaded dataframe shape:", feature_df.shape)

>>> generate_sample()
>>> pandas.__version__
'1.3.4'
>>> timeit.timeit(load_csv_file, number=1)
loaded dataframe shape: (5, 100000)
216.38200060278177

Installed Versions

INSTALLED VERSIONS

commit : 945c9ed
python : 3.8.10.final.0
python-bits : 64
OS : Linux
OS-release : 5.4.0-67-generic
Version : #75-Ubuntu SMP Fri Feb 19 18:03:38 UTC 2021
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.3.4
numpy : 1.21.2
pytz : 2021.3
dateutil : 2.8.2
pip : 21.3
setuptools : 44.0.0
Cython : 0.29.24
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 3.0.2
IPython : 7.28.0
pandas_datareader: None
bs4 : None
bottleneck : None
fsspec : None
fastparquet : None
gcsfs : None
matplotlib : 3.4.3
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pyxlsb : None
s3fs : None
scipy : 1.7.1
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlwt : None
numba : None

Prior Performance

import os
import pandas
import numpy
import timeit

def generate_sample():
    if os.path.exists("test_small.csv.gz") == False:
        nb_col = 100000
        nb_row = 5
        feature_list = {'sample': ['s_' + str(i+1) for i in range(nb_row)]}
        for i in range(nb_col):
            feature_list.update({'feature_' + str(i+1): list(numpy.random.uniform(low=0, high=10, size=nb_row))})
        df = pandas.DataFrame(feature_list)
        df.to_csv("test_small.csv.gz", index=False, float_format="%.6f")

def load_csv_file():
    col_names = pandas.read_csv("test_small.csv.gz", low_memory=False, nrows=1).columns
    types_dict = {col: numpy.float32 for col in col_names}
    types_dict.update({'sample': str})
    feature_df = pandas.read_csv("test_small.csv.gz", index_col="sample", na_filter=False, dtype=types_dict, low_memory=False)
    print("loaded dataframe shape:", feature_df.shape)

>>> generate_sample()
>>> pandas.__version__
'1.2.5'
>>> timeit.timeit(load_csv_file, number=1)
loaded dataframe shape: (5, 100000)
6.001252222806215
@haghshenas haghshenas added Needs Triage Issue that has not been reviewed by a pandas team member Performance Memory or execution speed performance labels Oct 19, 2021
@jbrockmendel
Copy link
Member

Have you done any profiling to narrow down where the issue is? Does the issue depend on having a 1e5 columns or would it show up with e.g. 100? What if you used a BytesIO object instead of writing to a file?

@jbrockmendel
Copy link
Member

Looks like a lot of time is spent in _set_noconvert_columns that wasn't before

@haghshenas
Copy link
Author

haghshenas commented Oct 19, 2021 •
edited
Loading

I do see the same issue using BytesIO object instead of file.
The issue definitely depends on the number of columns. Here is my test using BytesIO object:

No. columns Speed of v1.2.5 vs. v1.3.4
1000 0.75
10000 2.9
100000 23.9

@jbrockmendel jbrockmendel mentioned this issue Oct 26, 2021
Merged
4 tasks
@jreback jreback added this to the 1.3.5 milestone Oct 26, 2021
@simonjayhawkins simonjayhawkins mentioned this issue Nov 10, 2021
Closed
3 tasks
This was referenced Jan 29, 2024
Open
Open
Open
This was referenced Feb 8, 2024
Open
Open
Open
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
Needs Triage Issue that has not been reviewed by a pandas team member Performance Memory or execution speed performance
Projects
None yet
Milestone
1.3.5
Development

Successfully merging a pull request may close this issue.

PERF: read_csv GH#44106 jbrockmendel/pandas
3 participants
@jreback @haghshenas @jbrockmendel