You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Sep 8, 2025. It is now read-only.
Repository navigation
This repository was archived by the owner on Sep 8, 2025. It is now read-only.
Currently load.py and dump.py utilities are loading/dumping documents one by one which is tremendously slow.
Introducing bulk loading/dumping will really speed up the things here.
Maybe we can add an option like "--bulk-size" with default value set to 1 (load/dump documents one by one, just like it happens now) to allow user some additional utility tuning.
Good stuff! For inclusion into CouchDB-Python, I have a number of requests:
Please remove the change in .hgignore, as it isn't needed anymore
Please see if you can add a test for the new behavior
It would be great if you can split this into two patches: one that abstracts writing into a separate function, and another one that actually does the bulk requests/writes -- this makes it easier to review the changes now and in the future
I've pushed modified versions; for r6f91fa675423 , I:
Renamed function from write_dump() to dump_doc()
Moved dump_doc() outside dump_db(), added envelope argument
Rewrote commit message to clarify
In re8cafe210d91 , I:
Made sure lines didn't get longer than 80 chars
Tightened up the loop code (while True, if condition: break is a little silly)
Rewrote commit message to clarify
Could you redo your bulk loading along these lines? You also introduce a bug wrt error handling; db.update() doesn't throw Exceptions like db.setattr(). Also, your test case references a test data file that isn't included in the patch.
Activity
djc commentedon Jul 12, 2014
From Pavel.Ts...@gmail.com on June 14, 2013 08:10:36
I'm working on initial implementation here, will provide some patches later
djc commentedon Jul 12, 2014
From Pavel.Ts...@gmail.com on June 17, 2013 05:10:15
I finished bulk dumping documents. You can see it here https://code.google.com/r/paveltsipinio-bulk-dumping/source/detail?r=e0f1bda24cc0bc487bf782ebdabc9d817bf7d4f6&name=bulk_dumping
djc commentedon Jul 12, 2014
From djc.ochtman on June 17, 2013 05:44:32
Good stuff! For inclusion into CouchDB-Python, I have a number of requests:
djc commentedon Jul 12, 2014
From Pavel.Ts...@gmail.com on June 17, 2013 10:23:15
I fixed your requests and added bulk load method. https://code.google.com/r/paveltsipinio-bulk-dumping/source/detail?r=de81adea330909f13d9bf37f98e25d4b7c657a92&name=bulk_dumping
djc commentedon Jul 12, 2014
From djc.ochtman on June 18, 2013 01:00:06
I've pushed modified versions; for r6f91fa675423 , I:
In re8cafe210d91 , I:
Could you redo your bulk loading along these lines? You also introduce a bug wrt error handling; db.update() doesn't throw Exceptions like db.setattr(). Also, your test case references a test data file that isn't included in the patch.
djc commentedon Jul 12, 2014
From Pavel.Ts...@gmail.com on June 18, 2013 08:32:57
I hope I clearly understand your recommendations about code design. I pushed it to https://code.google.com/r/paveltsipinio-bulk-dumping/source/detail?r=46b5043fe465274850c4a821e468ca9ca90b70e0&name=bulk_dumping I did not understand what you mean about test data file. I don't have any test data files.