Repository navigation
Reject oversized TEXTSPLIT results before building the padded grid (#538) - #620
Merged
Merged
Conversation
MarekSpodBiedry
marked this pull request as ready for review
October 12, 2026 06:54
Member
|
Merged — thank you @MarekSpodBiedry for the contribution! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixes #538.
TEXTSPLIT pads ragged rows to the widest row, reserved and filled the whole
nrows * ncolsgrid, and only then letarray_valreject it with#NUM!. A text under the 32,767-character cell limit (one 8,000-column row followed by 8,000 one-cell rows) asked for about 64 million cells before being rejected.The fix checks
nrows.saturating_mul(ncols) > MAX_CELLSright after the dimensions are known and returns#NUM!before any allocation. This is the same guard the sibling array builders already use (dynamic.rsMAKEARRAY/SEQUENCE/EXPAND,info.rs). Results within the limit, padding and the default#N/Afill are unchanged.Test:
split_rejects_oversized_padded_gridfeeds the 8,000 x 8,001 case and expects#NUM!. Without the fix the result is still#NUM!, but only after building the ~64M-cell grid, so the test is slow and memory-heavy rather than red.