From 30192d98024f343b050fe1e8e2ff49f2a456b0a3 Mon Sep 17 00:00:00 2001 From: Ruihang Xia Date: Mon, 10 Nov 2025 15:59:25 +0800 Subject: [PATCH] feat: disable default compression for `__op_type` column (#7196) * feat: disable default compression for `__op_type` column Signed-off-by: Ruihang Xia * update test Signed-off-by: Ruihang Xia * revert unrelated code Signed-off-by: Ruihang Xia --------- Signed-off-by: Ruihang Xia --- src/metric-engine/src/engine/flush.rs | 12 ++++++------ src/mito2/src/engine/basic_test.rs | 12 ++++++------ src/mito2/src/sst/parquet/writer.rs | 4 +++- 3 files changed, 15 insertions(+), 13 deletions(-) diff --git a/src/metric-engine/src/engine/flush.rs b/src/metric-engine/src/engine/flush.rs index a962243525..e7a2400a84 100644 --- a/src/metric-engine/src/engine/flush.rs +++ b/src/metric-engine/src/engine/flush.rs @@ -128,12 +128,12 @@ mod tests { assert_eq!( debug_format, r#" -ManifestSstEntry { table_dir: "test_metric_region/", region_id: 47244640257(11, 1), table_id: 11, region_number: 1, region_group: 0, region_sequence: 1, file_id: "", index_file_id: Some(""), level: 0, file_path: "test_metric_region/11_0000000001/data/.parquet", file_size: 3173, index_file_path: Some("test_metric_region/11_0000000001/data/index/.puffin"), index_file_size: Some(235), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9::Millisecond, sequence: Some(20), origin_region_id: 47244640257(11, 1), node_id: None, visible: true } -ManifestSstEntry { table_dir: "test_metric_region/", region_id: 47244640258(11, 2), table_id: 11, region_number: 2, region_group: 0, region_sequence: 2, file_id: "", index_file_id: Some(""), level: 0, file_path: "test_metric_region/11_0000000002/data/.parquet", file_size: 3173, index_file_path: Some("test_metric_region/11_0000000002/data/index/.puffin"), index_file_size: Some(235), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9::Millisecond, sequence: Some(10), origin_region_id: 47244640258(11, 2), node_id: None, visible: true } -ManifestSstEntry { table_dir: "test_metric_region/", region_id: 47261417473(11, 16777217), table_id: 11, region_number: 16777217, region_group: 1, region_sequence: 1, file_id: "", index_file_id: None, level: 0, file_path: "test_metric_region/11_0000000001/metadata/.parquet", file_size: 3505, index_file_path: None, index_file_size: None, num_rows: 8, num_row_groups: 1, num_series: Some(8), min_ts: 0::Millisecond, max_ts: 0::Millisecond, sequence: Some(8), origin_region_id: 47261417473(11, 16777217), node_id: None, visible: true } -ManifestSstEntry { table_dir: "test_metric_region/", region_id: 47261417474(11, 16777218), table_id: 11, region_number: 16777218, region_group: 1, region_sequence: 2, file_id: "", index_file_id: None, level: 0, file_path: "test_metric_region/11_0000000002/metadata/.parquet", file_size: 3489, index_file_path: None, index_file_size: None, num_rows: 4, num_row_groups: 1, num_series: Some(4), min_ts: 0::Millisecond, max_ts: 0::Millisecond, sequence: Some(4), origin_region_id: 47261417474(11, 16777218), node_id: None, visible: true } -ManifestSstEntry { table_dir: "test_metric_region/", region_id: 94489280554(22, 42), table_id: 22, region_number: 42, region_group: 0, region_sequence: 42, file_id: "", index_file_id: Some(""), level: 0, file_path: "test_metric_region/22_0000000042/data/.parquet", file_size: 3173, index_file_path: Some("test_metric_region/22_0000000042/data/index/.puffin"), index_file_size: Some(235), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9::Millisecond, sequence: Some(10), origin_region_id: 94489280554(22, 42), node_id: None, visible: true } -ManifestSstEntry { table_dir: "test_metric_region/", region_id: 94506057770(22, 16777258), table_id: 22, region_number: 16777258, region_group: 1, region_sequence: 42, file_id: "", index_file_id: None, level: 0, file_path: "test_metric_region/22_0000000042/metadata/.parquet", file_size: 3489, index_file_path: None, index_file_size: None, num_rows: 4, num_row_groups: 1, num_series: Some(4), min_ts: 0::Millisecond, max_ts: 0::Millisecond, sequence: Some(4), origin_region_id: 94506057770(22, 16777258), node_id: None, visible: true }"# +ManifestSstEntry { table_dir: "test_metric_region/", region_id: 47244640257(11, 1), table_id: 11, region_number: 1, region_group: 0, region_sequence: 1, file_id: "", index_file_id: Some(""), level: 0, file_path: "test_metric_region/11_0000000001/data/.parquet", file_size: 3155, index_file_path: Some("test_metric_region/11_0000000001/data/index/.puffin"), index_file_size: Some(235), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9::Millisecond, sequence: Some(20), origin_region_id: 47244640257(11, 1), node_id: None, visible: true } +ManifestSstEntry { table_dir: "test_metric_region/", region_id: 47244640258(11, 2), table_id: 11, region_number: 2, region_group: 0, region_sequence: 2, file_id: "", index_file_id: Some(""), level: 0, file_path: "test_metric_region/11_0000000002/data/.parquet", file_size: 3155, index_file_path: Some("test_metric_region/11_0000000002/data/index/.puffin"), index_file_size: Some(235), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9::Millisecond, sequence: Some(10), origin_region_id: 47244640258(11, 2), node_id: None, visible: true } +ManifestSstEntry { table_dir: "test_metric_region/", region_id: 47261417473(11, 16777217), table_id: 11, region_number: 16777217, region_group: 1, region_sequence: 1, file_id: "", index_file_id: None, level: 0, file_path: "test_metric_region/11_0000000001/metadata/.parquet", file_size: 3487, index_file_path: None, index_file_size: None, num_rows: 8, num_row_groups: 1, num_series: Some(8), min_ts: 0::Millisecond, max_ts: 0::Millisecond, sequence: Some(8), origin_region_id: 47261417473(11, 16777217), node_id: None, visible: true } +ManifestSstEntry { table_dir: "test_metric_region/", region_id: 47261417474(11, 16777218), table_id: 11, region_number: 16777218, region_group: 1, region_sequence: 2, file_id: "", index_file_id: None, level: 0, file_path: "test_metric_region/11_0000000002/metadata/.parquet", file_size: 3471, index_file_path: None, index_file_size: None, num_rows: 4, num_row_groups: 1, num_series: Some(4), min_ts: 0::Millisecond, max_ts: 0::Millisecond, sequence: Some(4), origin_region_id: 47261417474(11, 16777218), node_id: None, visible: true } +ManifestSstEntry { table_dir: "test_metric_region/", region_id: 94489280554(22, 42), table_id: 22, region_number: 42, region_group: 0, region_sequence: 42, file_id: "", index_file_id: Some(""), level: 0, file_path: "test_metric_region/22_0000000042/data/.parquet", file_size: 3155, index_file_path: Some("test_metric_region/22_0000000042/data/index/.puffin"), index_file_size: Some(235), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9::Millisecond, sequence: Some(10), origin_region_id: 94489280554(22, 42), node_id: None, visible: true } +ManifestSstEntry { table_dir: "test_metric_region/", region_id: 94506057770(22, 16777258), table_id: 22, region_number: 16777258, region_group: 1, region_sequence: 42, file_id: "", index_file_id: None, level: 0, file_path: "test_metric_region/22_0000000042/metadata/.parquet", file_size: 3471, index_file_path: None, index_file_size: None, num_rows: 4, num_row_groups: 1, num_series: Some(4), min_ts: 0::Millisecond, max_ts: 0::Millisecond, sequence: Some(4), origin_region_id: 94506057770(22, 16777258), node_id: None, visible: true }"# ); // list from storage let storage_entries = mito diff --git a/src/mito2/src/engine/basic_test.rs b/src/mito2/src/engine/basic_test.rs index bf0ba6449b..f17726abef 100644 --- a/src/mito2/src/engine/basic_test.rs +++ b/src/mito2/src/engine/basic_test.rs @@ -861,9 +861,9 @@ async fn test_cache_null_primary_key_with_format(flat_format: bool) { #[tokio::test] async fn test_list_ssts() { test_list_ssts_with_format(false, r#" -ManifestSstEntry { table_dir: "test/", region_id: 47244640257(11, 1), table_id: 11, region_number: 1, region_group: 0, region_sequence: 1, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/11_0000000001/.parquet", file_size: 2531, index_file_path: Some("test/11_0000000001/index/.puffin"), index_file_size: Some(250), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 47244640257(11, 1), node_id: None, visible: true } -ManifestSstEntry { table_dir: "test/", region_id: 47244640258(11, 2), table_id: 11, region_number: 2, region_group: 0, region_sequence: 2, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/11_0000000002/.parquet", file_size: 2531, index_file_path: Some("test/11_0000000002/index/.puffin"), index_file_size: Some(250), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 47244640258(11, 2), node_id: None, visible: true } -ManifestSstEntry { table_dir: "test/", region_id: 94489280554(22, 42), table_id: 22, region_number: 42, region_group: 0, region_sequence: 42, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/22_0000000042/.parquet", file_size: 2531, index_file_path: Some("test/22_0000000042/index/.puffin"), index_file_size: Some(250), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 94489280554(22, 42), node_id: None, visible: true }"# ,r#" +ManifestSstEntry { table_dir: "test/", region_id: 47244640257(11, 1), table_id: 11, region_number: 1, region_group: 0, region_sequence: 1, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/11_0000000001/.parquet", file_size: 2513, index_file_path: Some("test/11_0000000001/index/.puffin"), index_file_size: Some(250), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 47244640257(11, 1), node_id: None, visible: true } +ManifestSstEntry { table_dir: "test/", region_id: 47244640258(11, 2), table_id: 11, region_number: 2, region_group: 0, region_sequence: 2, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/11_0000000002/.parquet", file_size: 2513, index_file_path: Some("test/11_0000000002/index/.puffin"), index_file_size: Some(250), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 47244640258(11, 2), node_id: None, visible: true } +ManifestSstEntry { table_dir: "test/", region_id: 94489280554(22, 42), table_id: 22, region_number: 42, region_group: 0, region_sequence: 42, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/22_0000000042/.parquet", file_size: 2513, index_file_path: Some("test/22_0000000042/index/.puffin"), index_file_size: Some(250), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 94489280554(22, 42), node_id: None, visible: true }"# ,r#" StorageSstEntry { file_path: "test/11_0000000001/.parquet", file_size: None, last_modified_ms: None, node_id: None } StorageSstEntry { file_path: "test/11_0000000001/index/.puffin", file_size: None, last_modified_ms: None, node_id: None } StorageSstEntry { file_path: "test/11_0000000002/.parquet", file_size: None, last_modified_ms: None, node_id: None } @@ -871,9 +871,9 @@ StorageSstEntry { file_path: "test/11_0000000002/index/.puffin", file_s StorageSstEntry { file_path: "test/22_0000000042/.parquet", file_size: None, last_modified_ms: None, node_id: None } StorageSstEntry { file_path: "test/22_0000000042/index/.puffin", file_size: None, last_modified_ms: None, node_id: None }"#).await; test_list_ssts_with_format(true, r#" -ManifestSstEntry { table_dir: "test/", region_id: 47244640257(11, 1), table_id: 11, region_number: 1, region_group: 0, region_sequence: 1, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/11_0000000001/.parquet", file_size: 2855, index_file_path: Some("test/11_0000000001/index/.puffin"), index_file_size: Some(292), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 47244640257(11, 1), node_id: None, visible: true } -ManifestSstEntry { table_dir: "test/", region_id: 47244640258(11, 2), table_id: 11, region_number: 2, region_group: 0, region_sequence: 2, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/11_0000000002/.parquet", file_size: 2855, index_file_path: Some("test/11_0000000002/index/.puffin"), index_file_size: Some(292), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 47244640258(11, 2), node_id: None, visible: true } -ManifestSstEntry { table_dir: "test/", region_id: 94489280554(22, 42), table_id: 22, region_number: 42, region_group: 0, region_sequence: 42, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/22_0000000042/.parquet", file_size: 2855, index_file_path: Some("test/22_0000000042/index/.puffin"), index_file_size: Some(292), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 94489280554(22, 42), node_id: None, visible: true }"#, r#" +ManifestSstEntry { table_dir: "test/", region_id: 47244640257(11, 1), table_id: 11, region_number: 1, region_group: 0, region_sequence: 1, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/11_0000000001/.parquet", file_size: 2837, index_file_path: Some("test/11_0000000001/index/.puffin"), index_file_size: Some(292), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 47244640257(11, 1), node_id: None, visible: true } +ManifestSstEntry { table_dir: "test/", region_id: 47244640258(11, 2), table_id: 11, region_number: 2, region_group: 0, region_sequence: 2, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/11_0000000002/.parquet", file_size: 2837, index_file_path: Some("test/11_0000000002/index/.puffin"), index_file_size: Some(292), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 47244640258(11, 2), node_id: None, visible: true } +ManifestSstEntry { table_dir: "test/", region_id: 94489280554(22, 42), table_id: 22, region_number: 42, region_group: 0, region_sequence: 42, file_id: "", index_file_id: Some(""), level: 0, file_path: "test/22_0000000042/.parquet", file_size: 2837, index_file_path: Some("test/22_0000000042/index/.puffin"), index_file_size: Some(292), num_rows: 10, num_row_groups: 1, num_series: Some(1), min_ts: 0::Millisecond, max_ts: 9000::Millisecond, sequence: Some(10), origin_region_id: 94489280554(22, 42), node_id: None, visible: true }"#, r#" StorageSstEntry { file_path: "test/11_0000000001/.parquet", file_size: None, last_modified_ms: None, node_id: None } StorageSstEntry { file_path: "test/11_0000000001/index/.puffin", file_size: None, last_modified_ms: None, node_id: None } StorageSstEntry { file_path: "test/11_0000000002/.parquet", file_size: None, last_modified_ms: None, node_id: None } diff --git a/src/mito2/src/sst/parquet/writer.rs b/src/mito2/src/sst/parquet/writer.rs index 857ec08878..5247e2eec8 100644 --- a/src/mito2/src/sst/parquet/writer.rs +++ b/src/mito2/src/sst/parquet/writer.rs @@ -40,7 +40,7 @@ use parquet::schema::types::ColumnPath; use smallvec::smallvec; use snafu::ResultExt; use store_api::metadata::RegionMetadataRef; -use store_api::storage::consts::SEQUENCE_COLUMN_NAME; +use store_api::storage::consts::{OP_TYPE_COLUMN_NAME, SEQUENCE_COLUMN_NAME}; use store_api::storage::{FileId, SequenceNumber}; use tokio::io::AsyncWrite; use tokio_util::compat::{Compat, FuturesAsyncWriteCompatExt}; @@ -388,12 +388,14 @@ where .clone(), ]); let seq_col = ColumnPath::new(vec![SEQUENCE_COLUMN_NAME.to_string()]); + let op_type_col = ColumnPath::new(vec![OP_TYPE_COLUMN_NAME.to_string()]); builder .set_column_encoding(seq_col.clone(), Encoding::DELTA_BINARY_PACKED) .set_column_dictionary_enabled(seq_col, false) .set_column_encoding(ts_col.clone(), Encoding::DELTA_BINARY_PACKED) .set_column_dictionary_enabled(ts_col, false) + .set_column_compression(op_type_col, Compression::UNCOMPRESSED) } async fn write_next_batch(