Disable write_marks_for_substreams_in_compact_parts by default

Author: Avogar

src/Core/SettingsChangesHistory.cpp

@@ -733,7 +733,7 @@ const VersionToSettingsChangesMap & getMergeTreeSettingsChangesHistory()
     {
         addSettingsChanges(merge_tree_settings_changes_history, "25.5",
         {
-            {"write_marks_for_substreams_in_compact_parts", false, true, "New setting"},
+            {"write_marks_for_substreams_in_compact_parts", false, false, "New setting"},
         });
         addSettingsChanges(merge_tree_settings_changes_history, "25.4",
         {

src/Storages/MergeTree/MergeTreeSettings.cpp

@@ -252,7 +252,7 @@ namespace ErrorCodes
     This mode allows to use significantly less memory for storing discriminators
     in parts when there is mostly one variant or a lot of NULL values.
     )", 0) \
-    DECLARE(Bool, write_marks_for_substreams_in_compact_parts, true, R"(
+    DECLARE(Bool, write_marks_for_substreams_in_compact_parts, false, R"(
     Enables writing marks per each substream instead of per each column in Compact parts.
     It allows to read individual subcolumns from the data part efficiently.
     )", 0) \

tests/integration/test_replicated_merge_tree_s3/test.py

@@ -135,7 +135,7 @@ def drop_table(cluster):
 def test_insert_select_replicated(cluster, min_rows_for_wide_part, files_per_part):
     create_table(
         cluster,
-        additional_settings={"min_rows_for_wide_part": min_rows_for_wide_part},
+        additional_settings={"min_rows_for_wide_part": min_rows_for_wide_part, "write_marks_for_substreams_in_compact_parts": 1},
     )
 
     insert(cluster, node_idxs=[1, 2, 3], verify=True)

tests/integration/test_replicated_merge_tree_s3_zero_copy/test.py

@@ -116,7 +116,7 @@ def drop_table(cluster):
 def test_insert_select_replicated(cluster, min_rows_for_wide_part, files_per_part):
     create_table(
         cluster,
-        additional_settings="min_rows_for_wide_part={}".format(min_rows_for_wide_part),
+        additional_settings="min_rows_for_wide_part={}, write_marks_for_substreams_in_compact_parts={}".format(min_rows_for_wide_part, 1),
     )
 
     all_values = ""

tests/integration/test_s3_zero_copy_ttl/test.py

@@ -36,7 +36,7 @@ def test_ttl_move_and_s3(started_cluster):
             ORDER BY id
             PARTITION BY id
             TTL date TO DISK 's3_disk'
-            SETTINGS storage_policy='s3_and_default', temporary_directories_lifetime=1
+            SETTINGS storage_policy='s3_and_default', temporary_directories_lifetime=1, write_marks_for_substreams_in_compact_parts=1
             """.format(
                 i
             )

tests/queries/0_stateless/02675_profile_events_from_query_log_and_client.sh

@@ -52,7 +52,8 @@ CREATE TABLE times (t DateTime) ENGINE MergeTree ORDER BY t
     storage_policy='default',
     min_rows_for_wide_part = 1000000,
     min_bytes_for_wide_part = 1000000,
-    ratio_of_defaults_for_sparse_serialization=1.0;
+    ratio_of_defaults_for_sparse_serialization=1.0,
+    write_marks_for_substreams_in_compact_parts=1;
 "
 
 echo "INSERT"

tests/queries/0_stateless/02980_s3_plain_DROP_TABLE_MergeTree.sh

@@ -21,7 +21,7 @@ CLICKHOUSE_DATABASE="$new_database"
 
 $CLICKHOUSE_CLIENT -m -q "
     drop table if exists data;
-    create table data (key Int) engine=MergeTree() order by key;
+    create table data (key Int) engine=MergeTree() order by key settings write_marks_for_substreams_in_compact_parts=1;
     insert into data values (1);
     select 'data after INSERT', count() from data;
 "
@@ -34,6 +34,7 @@ $CLICKHOUSE_CLIENT -m -q "
     attach table data (key Int) engine=MergeTree() order by key
     settings
         max_suspicious_broken_parts=0,
+        write_marks_for_substreams_in_compact_parts=1,
         disk=disk(type=s3_plain,
             endpoint='http://localhost:11111/test/s3_plain/backups/$CLICKHOUSE_DATABASE',
             access_key_id='test',

tests/queries/0_stateless/02980_s3_plain_DROP_TABLE_ReplicatedMergeTree.sh

@@ -26,8 +26,8 @@ $CLICKHOUSE_CLIENT -m -q "
     drop table if exists data_read;
     drop table if exists data_write;
 
-    create table data_write (key Int) engine=ReplicatedMergeTree('/tables/{database}/data', 'write') order by key;
-    create table data_read (key Int) engine=ReplicatedMergeTree('/tables/{database}/data', 'read') order by key;
+    create table data_write (key Int) engine=ReplicatedMergeTree('/tables/{database}/data', 'write') order by key settings write_marks_for_substreams_in_compact_parts=1;
+    create table data_read (key Int) engine=ReplicatedMergeTree('/tables/{database}/data', 'read') order by key settings write_marks_for_substreams_in_compact_parts=1;
 
     insert into data_write values (1);
     system sync replica data_read;
@@ -42,6 +42,7 @@ $CLICKHOUSE_CLIENT -m -q "
     attach table data_read (key Int) engine=ReplicatedMergeTree('/tables/{database}/data', 'read') order by key
     settings
         max_suspicious_broken_parts=0,
+        write_marks_for_substreams_in_compact_parts=1,
         disk=disk(type=s3_plain,
             endpoint='http://localhost:11111/test/s3_plain/backups/$CLICKHOUSE_DATABASE',
             access_key_id='test',

tests/queries/0_stateless/03164_s3_settings_for_queries_and_merges.sql

@@ -10,7 +10,7 @@ SET s3_check_objects_after_upload=1;
 DROP TABLE IF EXISTS t_compact_bytes_s3;
 CREATE TABLE t_compact_bytes_s3(c1 UInt32, c2 UInt32, c3 UInt32, c4 UInt32, c5 UInt32)
 ENGINE = MergeTree ORDER BY c1
-SETTINGS index_granularity = 512, min_bytes_for_wide_part = '10G', storage_policy = 's3_no_cache';
+SETTINGS index_granularity = 512, min_bytes_for_wide_part = '10G', storage_policy = 's3_no_cache', write_marks_for_substreams_in_compact_parts=1;
 
 INSERT INTO t_compact_bytes_s3 SELECT number, number, number, number, number FROM numbers(512 * 32 * 40);