Function test_parquet_metadata_works

datafusion-cli/src/main.rs:510–540

()

    #[tokio::test]
    async fn test_parquet_metadata_works() -> Result<(), DataFusionError> {
        let ctx = SessionContext::new();
        ctx.register_udtf("parquet_metadata", Arc::new(ParquetMetadataFunc {}));

        // input with single quote
        let sql = "SELECT * FROM parquet_metadata('../datafusion/core/tests/data/fixed_size_list_array.parquet')";
        let df = ctx.sql(sql).await?;
        let rbs = df.collect().await?;

        assert_snapshot!(batches_to_string(&rbs), @r#"
        +-------------------------------------------------------------+--------------+--------------------+-----------------------+-----------------+-----------+-------------+------------+----------------+-------+-----------+-----------+------------------+----------------------+-----------------+-----------------+-------------+------------------------------+-------------------+------------------------+------------------+-----------------------+-------------------------+
        | filename                                                    | row_group_id | row_group_num_rows | row_group_num_columns | row_group_bytes | column_id | file_offset | num_values | path_in_schema | type  | stats_min | stats_max | stats_null_count | stats_distinct_count | stats_min_value | stats_max_value | compression | encodings                    | index_page_offset | dictionary_page_offset | data_page_offset | total_compressed_size | total_uncompressed_size |
        +-------------------------------------------------------------+--------------+--------------------+-----------------------+-----------------+-----------+-------------+------------+----------------+-------+-----------+-----------+------------------+----------------------+-----------------+-----------------+-------------+------------------------------+-------------------+------------------------+------------------+-----------------------+-------------------------+
        | ../datafusion/core/tests/data/fixed_size_list_array.parquet | 0            | 2                  | 1                     | 123             | 0         | 125         | 4          | "f0.list.item" | INT64 | 1         | 4         | 0                |                      | 1               | 4               | SNAPPY      | [PLAIN, RLE, RLE_DICTIONARY] |                   | 4                      | 46               | 121                   | 123                     |
        +-------------------------------------------------------------+--------------+--------------------+-----------------------+-----------------+-----------+-------------+------------+----------------+-------+-----------+-----------+------------------+----------------------+-----------------+-----------------+-------------+------------------------------+-------------------+------------------------+------------------+-----------------------+-------------------------+
        "#);

        // input with double quote
        let sql = "SELECT * FROM parquet_metadata(\"../datafusion/core/tests/data/fixed_size_list_array.parquet\")";
        let df = ctx.sql(sql).await?;
        let rbs = df.collect().await?;
        assert_snapshot!(batches_to_string(&rbs), @r#"
        +-------------------------------------------------------------+--------------+--------------------+-----------------------+-----------------+-----------+-------------+------------+----------------+-------+-----------+-----------+------------------+----------------------+-----------------+-----------------+-------------+------------------------------+-------------------+------------------------+------------------+-----------------------+-------------------------+
        | filename                                                    | row_group_id | row_group_num_rows | row_group_num_columns | row_group_bytes | column_id | file_offset | num_values | path_in_schema | type  | stats_min | stats_max | stats_null_count | stats_distinct_count | stats_min_value | stats_max_value | compression | encodings                    | index_page_offset | dictionary_page_offset | data_page_offset | total_compressed_size | total_uncompressed_size |
        +-------------------------------------------------------------+--------------+--------------------+-----------------------+-----------------+-----------+-------------+------------+----------------+-------+-----------+-----------+------------------+----------------------+-----------------+-----------------+-------------+------------------------------+-------------------+------------------------+------------------+-----------------------+-------------------------+
        | ../datafusion/core/tests/data/fixed_size_list_array.parquet | 0            | 2                  | 1                     | 123             | 0         | 125         | 4          | "f0.list.item" | INT64 | 1         | 4         | 0                |                      | 1               | 4               | SNAPPY      | [PLAIN, RLE, RLE_DICTIONARY] |                   | 4                      | 46               | 121                   | 123                     |
        +-------------------------------------------------------------+--------------+--------------------+-----------------------+-----------------+-----------+-------------+------------+----------------+-------+-----------+-----------+------------------+----------------------+-----------------+-----------------+-------------+------------------------------+-------------------+------------------------+------------------+-----------------------+-------------------------+
        "#);

        Ok(())
    }

    #[tokio::test]
    async fn test_parquet_metadata_works_with_strings() -> Result<(), DataFusionError> {