From 016d6653285225575fa692e915f3bf6dd07e5f16 Mon Sep 17 00:00:00 2001
From: Nikita Taranov <nikita.taranov@clickhouse.com>
Date: Tue, 28 May 2024 19:40:59 +0100
Subject: [PATCH] Revert "Remove 02151_hash_table_sizes_stats_distributed
 (fixes broken CI) (#54969)"

This reverts commit aa5c02c1be5e84c502b1b44c1164853d1f662199.
---
 ...sh_table_sizes_stats_distributed.reference | 33 +++++++
 ...2151_hash_table_sizes_stats_distributed.sh | 95 +++++++++++++++++++
 2 files changed, 128 insertions(+)
 create mode 100644 tests/queries/0_stateless/02151_hash_table_sizes_stats_distributed.reference
 create mode 100755 tests/queries/0_stateless/02151_hash_table_sizes_stats_distributed.sh

diff --git a/tests/queries/0_stateless/02151_hash_table_sizes_stats_distributed.reference b/tests/queries/0_stateless/02151_hash_table_sizes_stats_distributed.reference
new file mode 100644
index 00000000000..0d10114f4ff
--- /dev/null
+++ b/tests/queries/0_stateless/02151_hash_table_sizes_stats_distributed.reference
@@ -0,0 +1,33 @@
+1
+1
+--
+1
+1
+--
+1
+1
+--
+1
+1
+--
+1
+1
+1
+1
+--
+1
+1
+--
+1
+1
+1
+1
+--
+1
+1
+--
+1
+1
+1
+1
+--
diff --git a/tests/queries/0_stateless/02151_hash_table_sizes_stats_distributed.sh b/tests/queries/0_stateless/02151_hash_table_sizes_stats_distributed.sh
new file mode 100755
index 00000000000..77b9b2942c5
--- /dev/null
+++ b/tests/queries/0_stateless/02151_hash_table_sizes_stats_distributed.sh
@@ -0,0 +1,95 @@
+#!/usr/bin/env bash
+# Tags: long, distributed, no-tsan, no-debug
+
+# These tests don't use `current_database = currentDatabase()` condition, because database name isn't propagated during remote queries.
+
+# shellcheck disable=SC2154
+
+CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
+# shellcheck source=../shell_config.sh
+. "$CURDIR"/../shell_config.sh
+
+
+# tests rely on that all the rows are unique and max_threads divides table_size
+table_size=1000005
+max_threads=5
+
+
+prepare_table() {
+  table_name="t_hash_table_sizes_stats_$RANDOM$RANDOM"
+  $CLICKHOUSE_CLIENT -q "DROP TABLE IF EXISTS $table_name;"
+  if [ -z "$1" ]; then
+    $CLICKHOUSE_CLIENT -q "CREATE TABLE $table_name(number UInt64) Engine=MergeTree() ORDER BY tuple() SETTINGS index_granularity = 8192, index_granularity_bytes = '10Mi';"
+  else
+    $CLICKHOUSE_CLIENT -q "CREATE TABLE $table_name(number UInt64) Engine=MergeTree() ORDER BY $1 SETTINGS index_granularity = 8192, index_granularity_bytes = '10Mi';"
+  fi
+  $CLICKHOUSE_CLIENT -q "SYSTEM STOP MERGES $table_name;"
+  for ((i = 1; i <= max_threads; i++)); do
+    cnt=$((table_size / max_threads))
+    from=$(((i - 1) * cnt))
+    $CLICKHOUSE_CLIENT -q "INSERT INTO $table_name SELECT * FROM numbers($from, $cnt);"
+  done
+  $CLICKHOUSE_CLIENT -q "DROP TABLE IF EXISTS ${table_name}_d;"
+  $CLICKHOUSE_CLIENT -q "CREATE TABLE ${table_name}_d AS $table_name ENGINE = Distributed(test_cluster_two_shards, currentDatabase(), $table_name);"
+  table_name="${table_name}_d"
+}
+
+prepare_table_with_sorting_key() {
+  prepare_table "$1"
+}
+
+run_query() {
+  query_id="${CLICKHOUSE_DATABASE}_hash_table_sizes_stats_$RANDOM$RANDOM"
+  $CLICKHOUSE_CLIENT --query_id="$query_id" --multiquery -q "
+    SET max_block_size = $((table_size / 10));
+    SET merge_tree_min_rows_for_concurrent_read = 1;
+    SET max_untracked_memory = 0;
+    SET prefer_localhost_replica = 1;
+    $query"
+}
+
+check_preallocated_elements() {
+  $CLICKHOUSE_CLIENT -q "SYSTEM FLUSH LOGS"
+  # rows may be distributed in any way including "everything goes to the one particular thread"
+  min=$1
+  if [ -z "$2" ]; then
+    max=$1
+  else
+    max=$2
+  fi
+  $CLICKHOUSE_CLIENT --param_query_id="$query_id" -q "
+    SELECT COUNT(*)
+      FROM system.query_log
+     WHERE event_date >= yesterday() AND (query_id = {query_id:String} OR initial_query_id = {query_id:String})
+           AND ProfileEvents['AggregationPreallocatedElementsInHashTables'] BETWEEN $min AND $max
+  GROUP BY query_id"
+}
+
+check_convertion_to_two_level() {
+  $CLICKHOUSE_CLIENT -q "SYSTEM FLUSH LOGS"
+  # rows may be distributed in any way including "everything goes to the one particular thread"
+  $CLICKHOUSE_CLIENT --param_query_id="$query_id" -q "
+    SELECT SUM(ProfileEvents['AggregationHashTablesInitializedAsTwoLevel']) BETWEEN 1 AND $max_threads
+      FROM system.query_log
+     WHERE event_date >= yesterday() AND (query_id = {query_id:String} OR initial_query_id = {query_id:String})
+  GROUP BY query_id"
+}
+
+print_border() {
+  echo "--"
+}
+
+
+# shellcheck source=./02151_hash_table_sizes_stats.testcases
+source "$CURDIR"/02151_hash_table_sizes_stats.testcases
+
+
+test_one_thread_simple_group_by
+test_one_thread_simple_group_by_with_limit
+test_one_thread_simple_group_by_with_join_and_subquery
+test_several_threads_simple_group_by_with_limit_single_level_ht
+test_several_threads_simple_group_by_with_limit_two_level_ht
+test_several_threads_simple_group_by_with_limit_and_rollup_single_level_ht
+test_several_threads_simple_group_by_with_limit_and_rollup_two_level_ht
+test_several_threads_simple_group_by_with_limit_and_cube_single_level_ht
+test_several_threads_simple_group_by_with_limit_and_cube_two_level_ht