Map ClickHouse JSON type to PostgreSQL JSONB

doc/pg_clickhouse.md

@@ -310,7 +310,7 @@ types:
  Int32      | integer          | 
  Int64      | bigint           | 
  Int8       | smallint         | 
- JSON       | json             | HTTP engine only
+ JSON       | jsonb            | HTTP engine only
  String     | text             | 
  UInt16     | integer          | 
  UInt32     | bigint           | 
@@ -477,49 +477,50 @@ Set the `pg_clickhouse.session_settings` runtime parameter to configure
 [ClickHouse settings] to be set on subsequent queries. Example:
 
 ```sql
-SET pg_clickhouse.session_settings = 'join_use_nulls=1, final=1';
+SET pg_clickhouse.session_settings = 'join_use_nulls 1, final 1';
 ```
 
-The default is `join_use_nulls=1`. Set it to an empty string to fall back on
+The default is `join_use_nulls 1`. Set it to an empty string to fall back on
 the ClickHouse server's settings.
 
 ```sql
 SET pg_clickhouse.session_settings = '';
 ```
 
-The syntax is a comma-delimited list of key/value pairs separated by an equal
-sign. Keys must correspond to [ClickHouse settings]. Escape spaces, commas,
-and backslashes in values with a backslash:
+The syntax is a comma-delimited list of key/value pairs separated by one or
+more spaces. Keys must correspond to [ClickHouse settings]. Escape spaces,
+commas, and backslashes in values with a backslash:
 
 ```sql
-SET pg_clickhouse.session_settings = 'join_algorithm = grace_hash\,hash';
+SET pg_clickhouse.session_settings = 'join_algorithm grace_hash\,hash';
 ```
 
 Or use single quoted values to avoid escaping spaces and commas; consider
 using [dollar quoting] to avoid the need to double-quote:
 
 ```sql
-SET pg_clickhouse.session_settings = $$join_algorithm = 'grace_hash,hash'$$;
+SET pg_clickhouse.session_settings = $$join_algorithm 'grace_hash,hash'$$;
 ```
 
 If you care about legibility and need to set many settings, use multiple
 lines, for example:
 
 ```sql
 SET pg_clickhouse.session_settings TO $$
-    connect_timeout = 2,
-    count_distinct_implementation = uniq,
-    groupby_use_nulls = 0,
-    join_algorithm = 'prefer_partial_merge',
-    join_use_nulls = 1,
-    log_queries_min_type = QUERY_FINISH,
-    max_block_size = 32768,
-    max_execution_time = 45,
-    max_result_rows = 1024,
-    metrics_perf_events_list = 'this,that',
-    network_compression_method = ZSTD,
-    poll_interval = 5,
-    totals_mode = after_having_auto
+    connect_timeout 2,
+    count_distinct_implementation uniq,
+    final 1,
+    group_by_use_nulls 1,
+    join_algorithm 'prefer_partial_merge',
+    join_use_nulls 1,
+    log_queries_min_type QUERY_FINISH,
+    max_block_size 32768,
+    max_execution_time 45,
+    max_result_rows 1024,
+    metrics_perf_events_list 'this,that',
+    network_compression_method ZSTD,
+    poll_interval 5,
+    totals_mode after_having_auto
 $$;
 ```
 

src/connection.c

@@ -202,7 +202,7 @@ ch_connection_details *
 connstring_parse(const char *connstring)
 {
 	ListCell   *lc;
-	List	   *options = chfdw_parse_options(connstring, false);
+	List	   *options = chfdw_parse_options(connstring, false, true);
 	ch_connection_details *details = palloc0(sizeof(ch_connection_details));
 
 	if (options == NIL)

src/include/engine.h

@@ -24,6 +24,6 @@ typedef struct
 	const List	   *settings;
 }			ch_query;
 
-#define new_query(sql) {sql, chfdw_parse_options(ch_session_settings, true)}
+#define new_query(sql) {sql, chfdw_parse_options(ch_session_settings, true, false)}
 
 #endif							/* CLICKHOUSE_ENGINE_H */

src/include/fdw.h

@@ -203,7 +203,7 @@ extern char *ch_session_settings;
 extern void
 			chfdw_extract_options(List * defelems, char **driver, char **host, int *port,
 								  char **dbname, char **username, char **password);
-extern List * chfdw_parse_options(const char *options, bool with_comma);
+extern List * chfdw_parse_options(const char *options, bool with_comma, bool with_equal);
 
 /* in deparse.c */
 extern void chfdw_classify_conditions(PlannerInfo * root,

src/option.c

@@ -298,23 +298,34 @@ chfdw_extract_options(List * defelems, char **driver, char **host, int *port,
 /*
  * Parse options as key/value pairs. Used for connection parameters and
  * ClickHouse settings. Based on the Postgres conninfo_parse() function. The
- * format is:
+ * formats:
+ *
+ * with_comma = false, with_equal = false:
+ *
+ *     key = value key 'value'...
+ *
+ * with_comma = false, with_equal = true:
  *
  *     key = value key = 'value'...
  *
- * Each key/value pair must be comma-delimited if with_comma is true.
+ * with_comma = true, with_equal = false:
+ *
+ *     key value, key 'value',...
+ *
+ * with_comma = true, with_equal = true:
  *
  *     key = value, key = 'value',...
  *
- * Each key is an unquoted string followed by `=` with optional spaces
- * followed by the value. Values may contain backslash-escaped spaces,
- * backslashes, and commans when `with_comma` is true. Use SQL single-quoted
- * literals to remove the need to escape commas and spaces.
+ * Parameter names may not contain spaces or '=' (when with_equal is true),
+ * and support no escapes.
+ *
+ * Values may contain backslash-escaped spaces, backslashes, and commas. Use
+ * SQL single-quoted literals to remove the need to escape commas and spaces.
  *
  * Returns a PostgreSQL List containing DefElem cells.
  */
 List	   *
-chfdw_parse_options(const char *options_string, bool with_comma)
+chfdw_parse_options(const char *options_string, bool with_comma, bool with_equal)
 {
 	char	   *pname;
 	char	   *pval;
@@ -340,7 +351,7 @@ chfdw_parse_options(const char *options_string, bool with_comma)
 		pname = cp;
 		while (*cp)
 		{
-			if (*cp == '=')
+			if (with_equal && *cp == '=')
 				break;
 			if (isspace((unsigned char) *cp))
 			{
@@ -356,19 +367,18 @@ chfdw_parse_options(const char *options_string, bool with_comma)
 			cp++;
 		}
 
-		/* Check that there is a following '=' */
-		if (*cp != '=')
-			ereport(ERROR,
-					errcode(ERRCODE_SYNTAX_ERROR),
-					errmsg("pg_clickhouse: missing \"=\" after \"%s\" in options string", pname));
-		*cp++ = '\0';
-
-		/* Skip blanks after the '=' */
-		while (*cp)
+		if (with_equal)
 		{
-			if (!isspace((unsigned char) *cp))
-				break;
-			cp++;
+			/* Check that there is a following '=' */
+			if (*cp != '=')
+				ereport(ERROR,
+						errcode(ERRCODE_SYNTAX_ERROR),
+						errmsg("pg_clickhouse: missing \"=\" after \"%s\" in options string", pname));
+			*cp++ = '\0';
+
+			/* Skip blanks after the '=' */
+			while (isspace((unsigned char) *cp))
+				cp++;
 		}
 
 		/* Get the parameter value */
@@ -412,6 +422,10 @@ chfdw_parse_options(const char *options_string, bool with_comma)
 					*cp2++ = *cp++;
 			}
 			*cp2 = '\0';
+			if (cp2 == pval)
+				ereport(ERROR,
+						(errcode(ERRCODE_SYNTAX_ERROR),
+						 errmsg("pg_clickhouse: missing value for parameter \"%s\" in options string", pname)));
 		}
 		else
 		{
@@ -478,7 +492,7 @@ check_settings_guc(char **newval, void **extra, GucSource source)
 	/*
 	 * Make sure we can parse the settings.
 	 */
-	chfdw_parse_options(*newval, true);
+	chfdw_parse_options(*newval, true, false);
 
 	/*
 	 * All good if no error.
@@ -505,7 +519,7 @@ _PG_init(void)
 							   "Sets the default ClickHouse session settings.",
 							   NULL,
 							   &ch_session_settings,
-							   "join_use_nulls = 1, group_by_use_nulls = 1, final = 1",
+							   "join_use_nulls 1, group_by_use_nulls 1, final 1",
 							   PGC_USERSET,
 							   0,
 							   check_settings_guc,

src/pglink.c

@@ -400,6 +400,7 @@ extend_insert_query(ch_http_insert_state * state, TupleTableSlot * slot)
 				case VARCHAROID:
 				case TEXTOID:
 				case JSONOID:
+				case JSONBOID:
 				case NAMEOID:
 				case BITOID:
 				case BYTEAOID:
@@ -785,7 +786,7 @@ binary_insert_tuple(void *istate, TupleTableSlot * slot)
 		('UUID',     'uuid',             ''),
 		('IPv4',     'inet',             ''),
 		('IPv6',     'inet',             ''),
-		('JSON',     'json',             'HTTP engine only')
+		('JSON',     'jsonb',            'HTTP engine only')
 	) AS v("ClickHouse", "PostgreSQL", "Notes")
 	ORDER BY "ClickHouse";
 
@@ -810,7 +811,7 @@ static char *str_types_map[][2] = {
 	{"UUID", "UUID"},
 	{"IPv4", "inet"},
 	{"IPv6", "inet"},
-	{"JSON", "JSON"},
+	{"JSON", "JSONB"},
 	{NULL, NULL},
 };
 

test/expected/gucs.out

@@ -1,17 +1,12 @@
 \unset ECHO
- ch_noop_bigint 
-----------------
-
-(1 row)
-
 NOTICE:  OK ``
-NOTICE:  OK `join_use_nulls=1`
-NOTICE:  OK `join_use_nulls=1, xyz=true`
-NOTICE:  OK ` additional_result_filter = 'x != 2' `
-NOTICE:  OK ` additional_result_filter = 'x != 2' ,join_use_nulls = 1 `
-NOTICE:  OK ` xxx = DEFAULT, yyy = foo\,bar, zzz = 'He said, \'Hello\'', aaa = hi\ there `
-NOTICE:  ERR 42601 - pg_clickhouse: missing "=" after "join_use_nulls" in options string
-NOTICE:  ERR 42601 - pg_clickhouse: missing "=" after "join_use_nulls" in options string
+NOTICE:  OK `join_use_nulls 1`
+NOTICE:  OK `join_use_nulls 1, xyz true`
+NOTICE:  OK ` additional_result_filter 'x != 2' `
+NOTICE:  OK ` additional_result_filter 'x != 2' ,join_use_nulls 1 `
+NOTICE:  OK ` xxx DEFAULT, yyy foo\,bar, zzz    'He said, \'Hello\'', aaa  hi\ there `
+NOTICE:  ERR 42601 - pg_clickhouse: missing value for parameter "join_use_nulls" in options string
+NOTICE:  ERR 42601 - pg_clickhouse: missing comma after "join_use_nulls" value in options string
 NOTICE:  ERR 42601 - pg_clickhouse: unterminated quoted string in options string
 NOTICE:  ERR 42601 - pg_clickhouse: missing comma after "join_use_nulls" value in options string
         name        | value 
@@ -28,22 +23,22 @@ NOTICE:  ERR 42601 - pg_clickhouse: missing comma after "join_use_nulls" value i
  join_use_nulls     | 1
 (3 rows)
 
-        pg_clickhouse.session_settings        
-----------------------------------------------
-                                             +
-     connect_timeout = 2,                    +
-     count_distinct_implementation = uniq,   +
-     join_algorithm = 'prefer_partial_merge',+
-     join_use_nulls = 0,                     +
-     join_use_nulls = 1,                     +
-     log_queries_min_type = QUERY_FINISH,    +
-     max_block_size = 32768,                 +
-     max_execution_time = 45,                +
-     max_result_rows = 1024,                 +
-     metrics_perf_events_list = 'this,that', +
-     network_compression_method = ZSTD,      +
-     poll_interval = 5,                      +
-     totals_mode = after_having_auto         +
+       pg_clickhouse.session_settings       
+--------------------------------------------
+                                           +
+     connect_timeout 2,                    +
+     count_distinct_implementation uniq,   +
+     join_algorithm 'prefer_partial_merge',+
+     join_use_nulls 0,                     +
+     join_use_nulls 1,                     +
+     log_queries_min_type QUERY_FINISH,    +
+     max_block_size 32768,                 +
+     max_execution_time 45,                +
+     max_result_rows 1024,                 +
+     metrics_perf_events_list 'this,that', +
+     network_compression_method ZSTD,      +
+     poll_interval 5,                      +
+     totals_mode after_having_auto         +
 
 (1 row)
 

test/expected/json.out

@@ -34,7 +34,7 @@ IMPORT FOREIGN SCHEMA "json_test" FROM SERVER binary_json_loopback INTO json_bin
  Column |  Type   | Collation | Nullable | Default | FDW options 
 --------+---------+-----------+----------+---------+-------------
  id     | integer |           | not null |         | 
- data   | json    |           | not null |         | 
+ data   | jsonb   |           | not null |         | 
 Server: binary_json_loopback
 FDW options: (database 'json_test', table_name 'things', engine 'MergeTree')
 
@@ -44,7 +44,7 @@ IMPORT FOREIGN SCHEMA "json_test" FROM SERVER http_json_loopback INTO json_http;
  Column |  Type   | Collation | Nullable | Default | FDW options 
 --------+---------+-----------+----------+---------+-------------
  id     | integer |           | not null |         | 
- data   | json    |           | not null |         | 
+ data   | jsonb   |           | not null |         | 
 Server: http_json_loopback
 FDW options: (database 'json_test', table_name 'things', engine 'MergeTree')
 
@@ -64,12 +64,12 @@ SELECT * FROM json_bin.things ORDER BY id;
 ERROR:  pg_clickhouse: unsupported column type: JSON
 DETAIL:  Remote Query: SELECT id, data FROM json_test.things ORDER BY id ASC NULLS LAST
 SELECT * FROM json_http.things ORDER BY id;
- id |                           data                           
-----+----------------------------------------------------------
-  1 | {"id":1,"name":"widget","size":"large","stocked":true}
-  2 | {"id":2,"name":"sprocket","size":"small","stocked":true}
-  3 | {"id":3,"name":"gizmo","size":"medium","stocked":true}
-  4 | {"id":4,"name":"doodad","size":"large","stocked":false}
+ id |                              data                               
+----+-----------------------------------------------------------------
+  1 | {"id": 1, "name": "widget", "size": "large", "stocked": true}
+  2 | {"id": 2, "name": "sprocket", "size": "small", "stocked": true}
+  3 | {"id": 3, "name": "gizmo", "size": "medium", "stocked": true}
+  4 | {"id": 4, "name": "doodad", "size": "large", "stocked": false}
 (4 rows)
 
 SELECT clickhouse_raw_query('DROP DATABASE json_test');

test/expected/json_1.out

@@ -34,7 +34,7 @@ IMPORT FOREIGN SCHEMA "json_test" FROM SERVER binary_json_loopback INTO json_bin
  Column |  Type   | Collation | Nullable | Default | FDW options 
 --------+---------+-----------+----------+---------+-------------
  id     | integer |           | not null |         | 
- data   | json    |           | not null |         | 
+ data   | jsonb   |           | not null |         | 
 Server: binary_json_loopback
 FDW options: (database 'json_test', table_name 'things', engine 'MergeTree')
 
@@ -44,7 +44,7 @@ IMPORT FOREIGN SCHEMA "json_test" FROM SERVER http_json_loopback INTO json_http;
  Column |  Type   | Collation | Nullable | Default | FDW options 
 --------+---------+-----------+----------+---------+-------------
  id     | integer |           | not null |         | 
- data   | json    |           | not null |         | 
+ data   | jsonb   |           | not null |         | 
 Server: http_json_loopback
 FDW options: (database 'json_test', table_name 'things', engine 'MergeTree')
 
@@ -64,12 +64,12 @@ SELECT * FROM json_bin.things ORDER BY id;
 ERROR:  pg_clickhouse: unsupported column type: JSON
 DETAIL:  Remote Query: SELECT id, data FROM json_test.things ORDER BY id ASC NULLS LAST
 SELECT * FROM json_http.things ORDER BY id;
- id |                            data                            
-----+------------------------------------------------------------
-  1 | {"id":"1","name":"widget","size":"large","stocked":true}
-  2 | {"id":"2","name":"sprocket","size":"small","stocked":true}
-  3 | {"id":"3","name":"gizmo","size":"medium","stocked":true}
-  4 | {"id":"4","name":"doodad","size":"large","stocked":false}
+ id |                               data                                
+----+-------------------------------------------------------------------
+  1 | {"id": "1", "name": "widget", "size": "large", "stocked": true}
+  2 | {"id": "2", "name": "sprocket", "size": "small", "stocked": true}
+  3 | {"id": "3", "name": "gizmo", "size": "medium", "stocked": true}
+  4 | {"id": "4", "name": "doodad", "size": "large", "stocked": false}
 (4 rows)
 
 SELECT clickhouse_raw_query('DROP DATABASE json_test');