Add compressed table size to existing views

Some information views report hypertable sizes. Include
compressed table size in the calculation when applicable.

sql/views.sql

@@ -6,7 +6,8 @@ CREATE SCHEMA IF NOT EXISTS timescaledb_information;
 
 -- Convenience view to list all hypertables and their space usage
 CREATE OR REPLACE VIEW timescaledb_information.hypertable AS
-  SELECT ht.schema_name AS table_schema,
+WITH ht_size as (
+  SELECT ht.id, ht.schema_name AS table_schema,
     ht.table_name,
     t.tableowner AS table_owner,
     ht.num_dimensions,
@@ -14,15 +15,36 @@ CREATE OR REPLACE VIEW timescaledb_information.hypertable AS
      FROM _timescaledb_catalog.chunk ch
      WHERE ch.hypertable_id=ht.id
     ) AS num_chunks,
-    size.table_size,
+    bsize.table_bytes,
-    size.index_size,
+    bsize.index_bytes,
-    size.toast_size,
+    bsize.toast_bytes,
-    size.total_size
+    bsize.total_bytes
   FROM _timescaledb_catalog.hypertable ht
     LEFT OUTER JOIN pg_tables t ON ht.table_name=t.tablename AND ht.schema_name=t.schemaname
-    LEFT OUTER JOIN LATERAL @extschema@.hypertable_relation_size_pretty(
+    LEFT OUTER JOIN LATERAL @extschema@.hypertable_relation_size(
       CASE WHEN has_schema_privilege(ht.schema_name,'USAGE') THEN format('%I.%I',ht.schema_name,ht.table_name) ELSE NULL END
-    ) size ON true;
+    ) bsize ON true
+),
+compht_size as
+(
+  select srcht.id,
+  sum(map.compressed_heap_size) as heap_bytes,
+  sum(map.compressed_index_size) as index_bytes,
+  sum(map.compressed_toast_size) as toast_bytes,
+  sum(map.compressed_heap_size) + sum(map.compressed_toast_size) + sum(map.compressed_index_size) as total_bytes
+ FROM _timescaledb_catalog.chunk srcch, _timescaledb_catalog.compression_chunk_size map,
+      _timescaledb_catalog.hypertable srcht
+ where map.chunk_id = srcch.id and srcht.id = srcch.hypertable_id
+ group by srcht.id
+)
+select hts.table_schema, hts.table_name, hts.table_owner, 
+       hts.num_dimensions, hts.num_chunks,
+       pg_size_pretty( COALESCE(hts.table_bytes + compht_size.heap_bytes, hts.table_bytes)) as table_size,
+       pg_size_pretty( COALESCE(hts.index_bytes + compht_size.index_bytes , hts.index_bytes, compht_size.index_bytes)) as index_size,
+       pg_size_pretty( COALESCE(hts.toast_bytes + compht_size.toast_bytes, hts.toast_bytes, compht_size.toast_bytes)) as toast_size,
+       pg_size_pretty( COALESCE(hts.total_bytes + compht_size.total_bytes, hts.total_bytes)) as total_size
+FROM ht_size hts LEFT OUTER JOIN compht_size 
+ON hts.id = compht_size.id;
 
 CREATE OR REPLACE VIEW timescaledb_information.license AS
   SELECT _timescaledb_internal.license_edition() as edition,

tsl/test/expected/compression.out

@@ -393,6 +393,35 @@ compressed_index_bytes   | 32 kB
 compressed_toast_bytes   | 16 kB
 compressed_total_bytes   | 64 kB
 
+vacuum full foo;
+vacuum full conditions;
+-- After vacuum, table_bytes is 0, but any associated index/toast storage is not
+-- completely reclaimed. Sets it at 8K (page size). So a chunk which has
+-- been compressed still incurs an overhead of n * 8KB (for every index + toast table) storage on the original uncompressed chunk.
+select * from timescaledb_information.hypertable
+where table_name like 'foo' or table_name like 'conditions'
+order by table_name;
+-[ RECORD 1 ]--+------------------
+table_schema   | public
+table_name     | conditions
+table_owner    | default_perm_user
+num_dimensions | 1
+num_chunks     | 2
+table_size     | 16 kB
+index_size     | 48 kB
+toast_size     | 32 kB
+total_size     | 96 kB
+-[ RECORD 2 ]--+------------------
+table_schema   | public
+table_name     | foo
+table_owner    | default_perm_user
+num_dimensions | 1
+num_chunks     | 4
+table_size     | 32 kB
+index_size     | 144 kB
+toast_size     | 8192 bytes
+total_size     | 184 kB
+
 \x
 select decompress_chunk(ch1.schema_name|| '.' || ch1.table_name)
 FROM _timescaledb_catalog.chunk ch1, _timescaledb_catalog.hypertable ht where ch1.hypertable_id = ht.id and ht.table_name like 'conditions';

tsl/test/sql/compression.sql

@@ -155,6 +155,14 @@ where hypertable_name::text like 'conditions'
 order by hypertable_name, chunk_name;
 select * from timescaledb_information.compressed_hypertable_stats
 order by hypertable_name;
+vacuum full foo;
+vacuum full conditions;
+-- After vacuum, table_bytes is 0, but any associated index/toast storage is not
+-- completely reclaimed. Sets it at 8K (page size). So a chunk which has
+-- been compressed still incurs an overhead of n * 8KB (for every index + toast table) storage on the original uncompressed chunk.
+select * from timescaledb_information.hypertable
+where table_name like 'foo' or table_name like 'conditions'
+order by table_name;
 \x
 
 select decompress_chunk(ch1.schema_name|| '.' || ch1.table_name)