apache / apache/arrow-java

Bug Report: Column Nullability Not Honored in FlightSqlProducer Schema

Open
#782 0 comments 0 reactions 0 assignees View on GitHub
Type: bug
Dominant language
Java
Stars
94
Forks
152
Avg merge
3d 16h
Merged PRs (30d)
11

Description

### Describe the bug, including details regarding any error messages, version, and platform.

In the FlightSqlProducer class, the getFlightInfoPreparedStatement method is responsible for returning the Arrow Schema of the ResultSet for a given query. However, the returned schema does not accurately reflect the nullability of columns. Regardless of whether a column is defined as nullable or non-nullable, the schema always returns a fixed nullability value.

This leads to incorrect ResultSetMetaData when the schema is consumed via the Flight SQL driver.

Root Cause:

The issue originates in the method:

`org.apache.arrow.driver.jdbc.utils.ConvertUtils.convertArrowFieldsToColumnMetaDataList()`

Here, a list of `org.apache.arrow.vector.types.pojo.Field` objects is converted into `org.apache.calcite.avatica.ColumnMetaData` instances. However, the `setNullable` property of the `ColumnMetaData.Builder` is not being set based on the actual nullability of the Arrow field.

Expected Behavior:

The `setNullable` method of `ColumnMetaData.Builder` should be updated to reflect the value of `field.isNullable()` from the Arrow schema.

Impact:

This bug results in incorrect metadata being exposed to clients using the Flight SQL driver, potentially leading to misinterpretation of the data schema.

Contributor guide

Open the contributing guide

Research direction

Start with org.apache.arrow.driver.jdbc.utils.ConvertUtils.convertArrowFieldsToColumnMetaDataList(), which converts Arrow Field objects into ColumnMetaData instances for the Flight SQL driver. Check how the ColumnMetaData.Builder is populated and verify that the resulting metadata preserves each field's nullability. Done means nullable and non-nullable Arrow fields expose matching ResultSetMetaData values.

Written by the indexing model from the issue text.

Assessment

Tech stack
java
Domain
databases
Issue type
Bug
Difficulty
2/5
Estimated time
1-3 hours
Activity status
Stale
Clarity
Clearly specified
Newbie friendliness
55/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.