Is there an existing issue for this?

• I have searched the existing issues

Category of Bug / Issue

Reconcile bug

Current Behavior

When reconciling data from external sources via Lakehouse Federation (e.g., Teradata), the reconcile process fails with DELTA_EXCEED_CHAR_VARCHAR_LIMIT when writing intermediate DataFrames to Delta volumes.

Teradata CHAR(n) columns contain space-padded values. The intermediate Delta table/volume preserves the VARCHAR(n) type constraint from the source schema, causing the write to fail when padded values exceed the declared length.

Version: v0.12.2

Expected Behavior

Reconcile intermediate data should be written successfully regardless of source column padding. The intermediate tables/volumes are temporary and should use STRING (unbounded) instead of VARCHAR(n)/CHAR(n) to avoid length constraint violations.

Steps To Reproduce

1. Set up Lakehouse Federation connection to Teradata
2. Create a Foreign Catalog pointing to a Teradata database with CHAR(n) columns
3. Run Lakebridge Reconcile with data_source="databricks" using the Foreign Catalog as source
4. Source table has CHAR(16) columns with space-padded values
5. Reconcile fails when writing intermediate DataFrame to Volume

Relevant log output or Exception details

```
[DELTA_EXCEED_CHAR_VARCHAR_LIMIT] Value "4297690382031830 " exceeds char/varchar type length limitation.
Failed check: ((src_kyf IS NULL) OR (length(src_kyf) <= 16)).

File: recon_capture.py, line 92, in _write_df_to_volumes
df.write.format(self._format).save(path)
```

Operating System

macOS

Version

latest via Databricks CLI