Github user yucai commented on the issue:
https://github.com/apache/spark/pull/22197
@cloud-fan I also think my way changes too much in this PR.
> go through the parquet schema and find duplicated field names
If user query only query non-duplicated field, this way also throws
Exception. Like below:
```
spark.range(10).selectExpr("id as a", "id as b", "id as
B").write.mode("overwrite").parquet("/tmp/data")
sql("select a from t").collect
```
Is it acceptable? Or use another PR to do the refactor first?
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
