From 51ea0888e58edfdf0bfe5d233ff2222da21baf44 Mon Sep 17 00:00:00 2001
From: cangyin <excangyin@gmail.com>
Date: Fri, 17 Mar 2023 10:00:50 +0800
Subject: [PATCH] fix invalid segment id bug after mutation

skip hardlinking inverted index files in mutation
---
 src/Storages/MergeTree/MutateTask.cpp         | 10 ++++++++
 .../02346_inverted_index_mutation.reference   |  3 +++
 .../02346_inverted_index_mutation.sql         | 23 +++++++++++++++++++
 3 files changed, 36 insertions(+)
 create mode 100644 tests/queries/0_stateless/02346_inverted_index_mutation.reference
 create mode 100644 tests/queries/0_stateless/02346_inverted_index_mutation.sql

diff --git a/src/Storages/MergeTree/MutateTask.cpp b/src/Storages/MergeTree/MutateTask.cpp
index 9f7a12745c6..30fa50e325f 100644
--- a/src/Storages/MergeTree/MutateTask.cpp
+++ b/src/Storages/MergeTree/MutateTask.cpp
@@ -19,6 +19,7 @@
 #include <Storages/MergeTree/MergeTreeDataWriter.h>
 #include <Storages/MutationCommands.h>
 #include <Storages/MergeTree/MergeTreeDataMergerMutator.h>
+#include <Storages/MergeTree/MergeTreeIndexInverted.h>
 #include <DataTypes/DataTypeNullable.h>
 #include <boost/algorithm/string/replace.hpp>
 #include <Common/ProfileEventsScope.h>
@@ -551,6 +552,15 @@ static NameSet collectFilesToSkip(
         /// Since MinMax index has .idx2 extension, we need to add correct extension.
         files_to_skip.insert(index->getFileName() + index->getSerializedFileExtension());
         files_to_skip.insert(index->getFileName() + mrk_extension);
+
+        // skip all inverted index files, for they will be re-built
+        if (dynamic_cast<const MergeTreeIndexInverted *>(&*index) != nullptr)
+        {
+            files_to_skip.insert(index->getFileName() + ".gin_dict");
+            files_to_skip.insert(index->getFileName() + ".gin_post");
+            files_to_skip.insert(index->getFileName() + ".gin_sed");
+            files_to_skip.insert(index->getFileName() + ".gin_sid");
+        }
     }
 
     for (const auto & projection : projections_to_recalc)
diff --git a/tests/queries/0_stateless/02346_inverted_index_mutation.reference b/tests/queries/0_stateless/02346_inverted_index_mutation.reference
new file mode 100644
index 00000000000..b80f66e4c05
--- /dev/null
+++ b/tests/queries/0_stateless/02346_inverted_index_mutation.reference
@@ -0,0 +1,3 @@
+1
+2
+I am not inverted
diff --git a/tests/queries/0_stateless/02346_inverted_index_mutation.sql b/tests/queries/0_stateless/02346_inverted_index_mutation.sql
new file mode 100644
index 00000000000..0987b4aa43c
--- /dev/null
+++ b/tests/queries/0_stateless/02346_inverted_index_mutation.sql
@@ -0,0 +1,23 @@
+SET allow_experimental_inverted_index=1;
+
+DROP TABLE IF EXISTS t;
+CREATE TABLE t
+(
+    `timestamp` UInt64,
+    `s` String,
+    INDEX idx s TYPE inverted(3) GRANULARITY 1
+)
+ENGINE = MergeTree
+ORDER BY tuple()
+SETTINGS min_rows_for_wide_part = 1, min_bytes_for_wide_part = 1;
+
+INSERT INTO t (s) VALUES ('I am inverted');
+
+SELECT data_version FROM system.parts WHERE database=currentDatabase() AND table='t' AND active=1;
+
+-- do update column synchronously
+ALTER TABLE t UPDATE s='I am not inverted' WHERE 1 SETTINGS mutations_sync=1;
+
+SELECT data_version FROM system.parts WHERE database=currentDatabase() AND table='t' AND active=1;
+
+SELECT s FROM t WHERE s LIKE '%inverted%' SETTINGS force_data_skipping_indices='idx';