Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Don't load empty cells; use empirical worksheet dimensions #248
There can be cells that look empty to the human eye, but that are styled, e.g. someone applied a custom number format to the cell at some point. I propose we do not load those anymore. There's nothing to learn from their XML.
If such a cell still falls into the target rectangle, it will simply be
If such a cell falls outside the target rectangle, this PR prevents the creation of trailing row(s) or column(s) consisting entirely of
In addition to rows/columns of
To fully fix this, you also can't trust a worksheet's declared dimensions, because these cells "count". Therefore I propose we always compute dimension ourselves.
I haven't dealt with xls yet. I can leave this open and work on that. Or we could resolve this for xlsx and I'd open an issue to fix it for xls.
This branch builds off #247. I'm assuming it will be merged first.