Datafusion CLI crashes on large queries

[comphead]

Describe the bug

The formatting fails on large resultsets, presumably because of eagerly consuming the entire output, not just small rows set which are necessary for output

thread 'main' (29539648) panicked at datafusion-cli/src/print_format.rs:100:23:
Formatting argument out of range

To Reproduce

Have a large parquet file, then

> CREATE EXTERNAL TABLE test_table
                 STORED AS PARQUET
                 LOCATION '/tmp/generated_data.parquet_p1';

select * from test_table

Expected behavior

should not crash

Additional context

No response

[comphead]

I'll check it later unless someone volunteers

[Geethapranay1]

Hi @comphead , I looked into this.

The panic comes from keep_only_maxrows in print_format.rs (100). The format string uses {:<spaces$} which rust treats as a positional argument index. When the table is wide (many columns), the spaces value exceeds the number of format arguments and panics.

The fix replaces the format string with " ".repeat(spaces). This works for any table width.

Data collection in exec.rs already stops storing batches at the maxrows limit (default 40). The crash is in the formatting step, not eager consumption.

I have a fix with a unit test and a CLI integration test. Ready to open a PR. Let me know if you want anything else addressed here.