Truthound · Schema and type contracts · 5/6
Preserve field names, types and key semantics across versions.
119 articles
Polars is_in pushdown: checking filter meaning for nulls and negative zero
Compare inclusion decisions for original rows when checking filter equivalence across execution backends.
Duplicate Polars Map keys: when lookup and dictionary conversion choose different values
Define key uniqueness within each parent row and an explicit duplicate policy before consumers choose different Map operations.
Checking empty strings versus NULL across Polars CSV engines
Matching String dtypes do not establish that a quoted empty field retained its meaning.
Preserving parent validity when Polars Array extraction exposes zero
A child buffer value is not a valid measurement when its parent array row is NULL.
Iceberg upserts after partition evolution: check updates and deletes of existing keys
Successful writes to a new partition spec do not establish that rows written under the old spec were removed.
Before comparing Iceberg Evaluator and SQL NULL filters, choose the policy
If APIs intentionally use different NULL semantics, a result difference alone does not establish an error.
Check numeric text contracts before changing a DuckDB UNION
A UNION change can affect serialized numeric text as well as duplicate counts.
DuckDB WITH RECURSIVE and NULL branches: make the Boolean schema explicit
A branch containing only NULL still needs a clear type contract for downstream consumers.
Uppercase Doris rowset hex IDs: check collisions, not just parsing success
An identifier defect in empty rowsets warrants a contract check without being overstated as observed data loss.
NULLs after a DataFusion Parquet Variant merge: check presence by layout
Distinguish a field missing in the source from a reader-created NULL caused by representation handling.
DataFusion UNION ALL: validate declared schemas and executed batches separately
At export time, check that the schema observed in planning matches the type contract of the batch actually received.
Before a DataFusion Decimal comparison, check the integer-digit budget
A large precision value alone does not mean a promoted type has room for the required integer part.
Reused ClickHouse–Iceberg column names: compare filters by field ID
A familiar name can attach a current predicate to statistics belonging to a different historical field.
Iceberg binary truncate counts bytes: checking ClickHouse partition prefixes
Bytes resembling UTF-8 characters still require byte-based truncation in a binary partition.
Pydantic validate_call: making dropped TypedDict options visible to operators
Share the requested and accepted option sets, not just their values, so a dropped setting is not mistaken for a normal execution.
Pydantic TypeAdapter unions: checking the selected response branch before serialization
A type declaration alone does not prove that field projection occurred. Verify the selected branch and the paths actually emitted.
Pydantic exclude_none: teaching the difference between required keys and nullable values
Required describes key presence; nullable describes permitted values. The team's acceptance checklist must separate them.
Pydantic MISSING unions: checking that a parent response does not disclose child fields
Validate the public field set independently of the runtime object's available information. That keeps the agreed disclosure boundary visible.
Pydantic strict Literal: preserving the original JSON numeric kind
If a protocol distinguishes original numeric kinds, a normalized value alone cannot prove that the received representation met the contract.
Pydantic Enum dictionary keys: agreeing on the JSON wire mapping
In-memory key types differ from JSON object keys. Share an explicit mapping in both directions.
