Skip to main content

A fast and memory efficient way to load large CSV files (Timeseries data) into Pandas

Project description

fast_csv_loader.py

The csv_loader function efficiently loads a partial portion of a large CSV file containing time-series data into a pandas DataFrame.

The function allows:

  • Loading the last N lines from the end of the file.
  • Loading the last N lines from a specific date.

It can load any type of time-series (both timezone aware and Naive) and daily or intraday data.

It is useful for loading large datasets that may not fit entirely into memory. It also improves program execution time, when iterating or loading a large number of CSV files.

Supports Python >= 3.8

Install

pip install fast-csv-loader

Documentation

https://bennythadikaran.github.io/fast_csv_loader/

Performance

Loading a portion of a large file is significantly faster than loading the entire file in memory. Files used in the test were not particularly large. You may need to tweak the chunk_size parameter for your use case.

It is slower for smaller files or if you're loading nearly the entire portion of the file.

I chose a 6Kb chunk size based on testing with my specific requirements. Your requirements may differ.

csv_loader vs pandas.read_csv

Execution time - Last 160 lines

Execution time - Last 160 lines upto 1st Jan 2023

To run this performance test.

py tests/run.py

At the minimum, the CSV file must contain a Date and another column with newline chars at the end to correctly parse and load.

Date,Price\n
2023-12-01,200\n

Unit Test

To run the test:

py tests/test_csv_loader.py

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

fast_csv_loader-1.0.1.tar.gz (17.1 kB view hashes)

Uploaded Source

Built Distribution

fast_csv_loader-1.0.1-py3-none-any.whl (17.0 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page