Python client for OmniSci GPU-accelerated SQL engine and analytics platform
Clone or download
jp-harvey and randyzwitch Fix bug with loading Pandas dataframe row-wise (#133)
* Fixes for boolean na types, typos, na assignments in test cases (#110) (#3)

* Fix na types for boolean, fix test case na types, fillna prior to boolean Thrift cast to prevent type error

* Attempt to distinguish int from bigint when int column contains None, fix int64 test that was actually an in32 test, add int32 test case

* Fix linting issue

* Sync from source repo master (#4)

* Fixes for boolean na types, typos, na assignments in test cases (#110)

* Fix na types for boolean, fix test case na types, fillna prior to boolean Thrift cast to prevent type error

* Attempt to distinguish int from bigint when int column contains None, fix int64 test that was actually an in32 test, add int32 test case

* Fix linting issue

* Option to chunk Pandas columnar data load (#117)

* Fixes for boolean na types, typos, na assignments in test cases (#110) (#3)

* Fix na types for boolean, fix test case na types, fillna prior to boolean Thrift cast to prevent type error

* Attempt to distinguish int from bigint when int column contains None, fix int64 test that was actually an in32 test, add int32 test case

* Fix linting issue

* Add option for chunking a Pandas columnar data load

* Fix linting issues

* Reorder options

* Reorder options

* [WIP] initial changes for matching col types

* Convert to int for DATE types

* Revert "Convert to int for DATE types"

This reverts commit 04d4a6e.

* Revert "Revert "Convert to int for DATE types""

This reverts commit 4f2b2ee.

* Fix nulls and timestamp type

* Fix out of bounds date errors on date conversion

* Update date test to fall outside bounds of datetime64[ns]

* Convert numpy.ndarray to list

* Clarify confusing condition

* Fix fillna dictionary vaules for FLOAT and DOUBLE

* Remove cruft

* Nans fixes

* Add na type for DECIMAL

* Fix flake8 issues

* Add ability to load based on matching table column names to dataframe column names

* Add support for decimals

* Fix whitespace

* Remove extraneous whitespace

* Make load from table schema mandatory, dataframe column length validation

* Convert dataframe to array if loading rowwise
Latest commit 0453e9f Dec 5, 2018

README.rst

pymapd

https://travis-ci.org/omnisci/pymapd.svg?branch=master Documentation Status

A python DB API compliant interface for OmniSci (formerly MapD). See the documentation for more.

Quick Install

Packages are available on conda-forge and PyPI:

conda install -c conda-forge pymapd

Developer builds of each commit are available on the pymapd channel (Linux only):

conda install -c pymapd/label/dev -c conda-forge -c gpuopenanalytics/label/dev pymapd