feat(testdata): Let destinations bound generated 64-bit integers - #2554
Open
murarustefaan wants to merge 1 commit into
Open
feat(testdata): Let destinations bound generated 64-bit integers#2554murarustefaan wants to merge 1 commit into
murarustefaan wants to merge 1 commit into
Conversation
arrow-go v18.7.0 fixed large-integer precision in JSON decoding (apache/arrow-go#816). The test data generator builds records via builder.UnmarshalJSON, so until now every int64/uint64 it produced was silently rounded to a float64-representable value. They are now exact, and destinations whose storage or wire format round-trips integers through float64 cannot reproduce them. MaxIntegerBits lets those destinations declare the width they can represent exactly. Unset keeps the full 64-bit range, so consumers that do not opt in are unaffected. TableDiff now points at the option when a differing column matches after a float64 round-trip. Timestamp_ns already handles the same root cause unconditionally; the opt-in keeps real 64-bit coverage for destinations that are correct.
erezrokah
reviewed
Aug 3, 2026
erezrokah
left a comment
erezrokah
left a comment
Member
There was a problem hiding this comment.
There's some overlap with #2553 I think, @disq fixed this already in cloudquery/filetypes@b25d89f
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Current
arrow-go v18.7.0 fixed large-integer precision in JSON decoding (apache/arrow-go#816), and
TestDataGeneratorbuilds records viabuilder.UnmarshalJSON— so until now every int64/uint64 it produced was silently rounded to a float64-representable value. They are now exact, which turned 16 destination plugins red because their storage or wire format round-trips integers through float64.Expected
Destinations declare the integer width they can represent exactly via
MaxIntegerBits(schema.Float64SafeIntegerBitsfor the float64 case). Unset keeps the full 64-bit range, so consumers that do not opt in are unaffected, andTableDiffnow names the option when a differing column matches after a float64 round-trip.Why / Impact
Test-only API, additive, no behaviour change without opt-in;
timestamp_nsalready handles this same root cause unconditionally, and keeping this opt-in preserves real 64-bit coverage for destinations that are correct. Verified end-to-end: thefiledestination's csv failure at arrow-go v18.7.0 passes withMaxIntegerBits: 53, and removing any plumbing line now failsTestPluginFloat64Ints.🤖 Generated with Claude Code