[ARROW-4637] [Python] Avoid importing Pandas unless necessary

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 0.12.0
• Fix Version/s: 0.13.0
• Component/s: Python
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/21172
• Assignee: Wes McKinney
• Reporter: Antoine Pitrou
• Created: 20/Feb/19 12:26
• Updated: 11/Jan/23 07:35
• Resolved: 19/Mar/19 16:59

Description

Importing PyArrow is more than twice slower when Pandas is installed:

$ time python -c "import pyarrow"

real	0m0,360s
user	0m0,305s
sys	0m0,037s

$ time python -c "import sys; sys.modules['pandas'] = None; import pyarrow"

real	0m0,144s
user	0m0,124s
sys	0m0,020s

We should only import Pandas when necessary, e.g. when asked to ingest or create Pandas data.

Issue Links

• links to: GitHub Pull Request #3893
• links to: GitHub Pull Request #8588