o-spreadsheet icon indicating copy to clipboard operation
o-spreadsheet copied to clipboard

[PERF] table_conversion: avoid reading all cells multiple times

Open laa-odoo opened this issue 8 months ago • 1 comments

Previously, the content of each cell was inspected multiple times, once per table, leading to significant performance overhead. In this commit, the inspection process is optimized to traverse the cells only once.

A data structure now stores only the cells that need to be updated, improving both efficiency and calculation time.

This optimization results in a 10% reduction in total import time.

Task: 4699860

laa-odoo avatar Apr 03 '25 09:04 laa-odoo

Pull request status dashboard

robodoo avatar Apr 03 '25 09:04 robodoo

the commit title could point towards the global context in which it is used, which is the Xlsx import if i'm not. mistaken?

rrahir avatar Apr 15 '25 09:04 rrahir

the commit title could point towards the global context in which it is used, which is the Xlsx import if i'm not. mistaken?

in fact this concerns the xlsx import, the name of the commit has been changed

laa-odoo avatar Apr 17 '25 07:04 laa-odoo

Is it actually slower ? (for the common case where there are no table references) I exported the large formula dataset to re-import it. With this branch, convertTableFormulaReferences is 64ms, in 16.0 it's less than 1ms

should be ok

laa-odoo avatar Jun 24 '25 13:06 laa-odoo

robodoo r+

rrahir avatar Jul 03 '25 07:07 rrahir