How do I know that the bloom filter configuration is successful?

[madeirak]

Query engine

Spark=3.3.1
Iceberg=1.4.3

Question

After setting new table properties
'write.parquet.bloom-filter-enabled.column.xxx' = 'true',
'write.parquet.bloom-filter-max-bytes' = '100',
How do I know that the bloom filter configuration is successful?

For example, how to see from the parquet file

[ebyhr]

How about using parquet-cli?
footer option provides bloom filter's offset and length. Also, we can usebloom-filter option to check the column contains bloom filter or not, and whether the provided value doesn't exist or maybe exists in the file:

➜  parquet bloom-filter -c a -v 2 20250106_091836_00029_airuh-4a353534-10e4-4d7e-8d38-4b9edab7bf1b.parquet

Row group 0:
--------------------------------------------------------------------------------
column a has no bloom filter

➜  parquet bloom-filter -c b -v 1 20250106_091836_00029_airuh-4a353534-10e4-4d7e-8d38-4b9edab7bf1b.parquet

Row group 0:
--------------------------------------------------------------------------------
value 1 NOT exists.
➜  parquet bloom-filter -c b -v 2 20250106_091836_00029_airuh-4a353534-10e4-4d7e-8d38-4b9edab7bf1b.parquet

Row group 0:
--------------------------------------------------------------------------------
value 2 maybe exists.

[madeirak]

How about using parquet-cli? footer option provides bloom filter's offset and length. Also, we can usebloom-filter option to check the column contains bloom filter or not, and whether the provided value doesn't exist or maybe exists in the file:

➜  parquet bloom-filter -c a -v 2 20250106_091836_00029_airuh-4a353534-10e4-4d7e-8d38-4b9edab7bf1b.parquet

Row group 0:
--------------------------------------------------------------------------------
column a has no bloom filter

➜  parquet bloom-filter -c b -v 1 20250106_091836_00029_airuh-4a353534-10e4-4d7e-8d38-4b9edab7bf1b.parquet

Row group 0:
--------------------------------------------------------------------------------
value 1 NOT exists.
➜  parquet bloom-filter -c b -v 2 20250106_091836_00029_airuh-4a353534-10e4-4d7e-8d38-4b9edab7bf1b.parquet

Row group 0:
--------------------------------------------------------------------------------
value 2 maybe exists.

Thx for reply, It seems to work, I'll try it

[hashmapybx]

I'll try it also.the bloom filter is success.