Remove reindex_relation from recompression

We used to reindex relation when compressing chunks. Recently
we moved to inserting into indexes on compressed chunks in
order to reduce locks necessary for the operation. Since
recompression uses RowCompressor, it also started inserting
tuples into indexes but we never removed the relation reindexing.
This change removes the unnecessary reindex call.

.unreleased/pr_6529

@@ -0,0 +1 @@
+Implements: #6529 Remove reindex_relation from recompression

tsl/src/compression/api.c

@@ -1388,14 +1388,6 @@ tsl_recompress_chunk_segmentwise(PG_FUNCTION_ARGS)
 	index_close(index_rel, AccessExclusiveLock);
 	row_decompressor_close(&decompressor);
 
-#if PG14_LT
-	int options = 0;
-#else
-	ReindexParams params = { 0 };
-	ReindexParams *options = &params;
-#endif
-	reindex_relation(compressed_chunk->table_id, 0, options);
-
 	/* changed chunk status, so invalidate any plans involving this chunk */
 	CacheInvalidateRelcacheByRelid(uncompressed_chunk_id);
 	table_close(uncompressed_chunk_rel, ExclusiveLock);

tsl/test/sql/recompress_chunk_segmentwise.sql

@@ -211,6 +211,21 @@ call recompress_chunk(:'chunk_to_compress_mytab');
 select compressed_chunk_name as compressed_chunk_name_after_recompression from compressed_chunk_info_view where hypertable_name = 'mytab' \gset
 select :'compressed_chunk_name_before_recompression' as before_segmentwise_recompression, :'compressed_chunk_name_after_recompression' as after_segmentwise_recompression;
 
+-- delete will empty out the compressed chunk but keep its compressed status
+DELETE FROM mytab;
+
+INSERT INTO mytab 
+SELECT t, a, 3, 2 
+FROM generate_series('2023-01-01'::timestamptz, '2023-01-02'::timestamptz, '1 hour'::interval) t 
+CROSS JOIN generate_series(1, 10, 1) a;
+-- recompress will insert newly inserted tuples into compressed chunk along with inserting into the compressed chunk index
+CALL recompress_chunk(:'chunk_to_compress_mytab');
+-- make sure we are hitting the index and that the index contains the tuples
+SET enable_seqscan TO off;
+EXPLAIN (COSTS OFF) SELECT count(*) FROM mytab where a = 2;
+SELECT count(*) FROM mytab where a = 2;
+SET enable_seqscan TO default;
+
 SELECT decompress_chunk(show_chunks('mytab'));
 alter table mytab set (timescaledb.compress = false);
 alter table mytab set (timescaledb.compress);
@@ -241,7 +256,7 @@ select compressed_chunk_schema || '.' || compressed_chunk_name as compressed_chu
 call recompress_chunk(:'chunk_to_compress');
 
 select * from :compressed_chunk_name;
--- insert again, check both reindex works and NULL values properly handled
+-- insert again, check both index insertion works and NULL values properly handled
 insert into nullseg_one values (:'start_time', NULL, 4);
 call recompress_chunk(:'chunk_to_compress');
 select * from :compressed_chunk_name;
@@ -264,7 +279,7 @@ select compressed_chunk_schema || '.' || compressed_chunk_name as compressed_chu
 call recompress_chunk(:'chunk_to_compress');
 
 select * from :compressed_chunk_name;
--- insert again, check both reindex works and NULL values properly handled
+-- insert again, check both index insertion works and NULL values properly handled
 -- should match existing segment (1, NULL)
 insert into nullseg_many values (:'start_time', 1, NULL, NULL);
 call recompress_chunk(:'chunk_to_compress');