Add support for bit column types

BemiHQ · Dec 9, 2024 · 924a152 · 924a152
1 parent 8c42c30
commit 924a152
Show file tree

Hide file tree

Showing 7 changed files with 26 additions and 6 deletions.
diff --git a/README.md b/README.md
@@ -242,7 +242,7 @@ Primitive data types are mapped as follows:
 | PostgreSQL                                                  | Parquet                                           | Iceberg                          |
 |-------------------------------------------------------------|---------------------------------------------------|----------------------------------|
 | `bool`                                                      | `BOOLEAN`                                         | `boolean`                        |
-| `bpchar`, `varchar`, `text`                                 | `BYTE_ARRAY` (`UTF8`)                             | `string`                         |
+| `varchar`, `text`, `bpchar`, `bit`                          | `BYTE_ARRAY` (`UTF8`)                             | `string`                         |
 | `int2`, `int4`                                              | `INT32`                                           | `int`                            |
 | `int8`                                                      | `INT64`                                           | `long`                           |
 | `xid`                                                       | `INT32` (`UINT_32`)                               | `int`                            |

diff --git a/scripts/install.sh b/scripts/install.sh
@@ -1,6 +1,6 @@
 #!/bin/bash
 
-VERSION="0.19.4"
+VERSION="0.20.0"
 
 # Detect OS and architecture
 OS=$(uname -s | tr '[:upper:]' '[:lower:]')

diff --git a/scripts/test-data-types.sql b/scripts/test-data-types.sql
@@ -11,6 +11,7 @@ CREATE TYPE address AS (
 
 CREATE TABLE test_table (
   id SERIAL PRIMARY KEY,
+  bit_column BIT,
   bool_column BOOLEAN,
   bpchar_column BPCHAR(10),
   varchar_column VARCHAR(255),
@@ -48,6 +49,7 @@ CREATE TABLE test_table (
 );
 
 INSERT INTO test_table (
+  bit_column,
   bool_column,
   bpchar_column,
   varchar_column,
@@ -83,6 +85,7 @@ INSERT INTO test_table (
   array_ltree_column,
   user_defined_column
 ) VALUES (
+  B'1',                                     -- bit_column
   TRUE,                                     -- bool_column
   'bpchar',                                 -- bpchar_column
   'varchar',                                -- varchar_column
@@ -118,6 +121,7 @@ INSERT INTO test_table (
   '{"a.b", "c.d"}'::LTREE[],                -- array_ltree_column
   ROW('Toronto')                            -- user_defined_column
 ), (
+  NULL,                                     -- bit_column
   FALSE,                                    -- bool_column
   '',                                       -- bpchar_column
   NULL,                                     -- varchar_column

diff --git a/src/init_test.go b/src/init_test.go
@@ -6,6 +6,12 @@ import (
 )
 
 var TEST_PG_SCHEMA_COLUMNS = []PgSchemaColumn{
+	{
+		ColumnName:             "bit_column",
+		DataType:               "bit",
+		UdtName:                "bit",
+		CharacterMaximumLength: "1",
+	},
 	{
 		ColumnName: "bool_column",
 		DataType:   "boolean",
@@ -243,6 +249,7 @@ var TEST_PG_SCHEMA_COLUMNS = []PgSchemaColumn{
 
 var TEST_LOADED_ROWS = [][]string{
 	{
+		"1",                                    // bit_column
 		"true",                                 // bool_column
 		"bpchar",                               // bpchar_column
 		"varchar",                              // varchar_column
@@ -280,6 +287,7 @@ var TEST_LOADED_ROWS = [][]string{
 		"(Toronto)",                            // user_defined_column
 	},
 	{
+		PG_NULL_STRING,                     // bit_column
 		"false",                            // bool_column
 		"",                                 // bpchar_column
 		PG_NULL_STRING,                     // varchar_column

diff --git a/src/main.go b/src/main.go
@@ -6,7 +6,7 @@ import (
 	"time"
 )
 
-const VERSION = "0.19.4"
+const VERSION = "0.20.0"
 
 func main() {
 	config := LoadConfig()

diff --git a/src/pg_schema_column.go b/src/pg_schema_column.go
@@ -203,7 +203,7 @@ func (pgSchemaColumn *PgSchemaColumn) toParquetSchemaField() ParquetSchemaField
 
 func (pgSchemaColumn *PgSchemaColumn) parquetPrimitiveValue(value string) interface{} {
 	switch strings.TrimLeft(pgSchemaColumn.UdtName, "_") {
-	case "varchar", "char", "text", "bytea", "jsonb", "json", "numeric", "uuid", "interval",
+	case "varchar", "char", "text", "bit", "bytea", "jsonb", "json", "numeric", "uuid", "interval",
 		"point", "line", "lseg", "box", "path", "polygon", "circle",
 		"cidr", "inet", "macaddr", "macaddr8",
 		"tsvector", "pg_snapshot":
@@ -301,7 +301,7 @@ func (pgSchemaColumn *PgSchemaColumn) parquetPrimitiveValue(value string) interf
 
 func (pgSchemaColumn *PgSchemaColumn) parquetPrimitiveTypes() (primitiveType string, primitiveConvertedType string) {
 	switch strings.TrimLeft(pgSchemaColumn.UdtName, "_") {
-	case "varchar", "char", "text", "bpchar", "bytea", "interval", "jsonb", "json",
+	case "varchar", "char", "text", "bpchar", "bit", "bytea", "interval", "jsonb", "json",
 		"point", "line", "lseg", "box", "path", "polygon", "circle",
 		"cidr", "inet", "macaddr", "macaddr8",
 		"tsvector", "pg_snapshot":
@@ -350,7 +350,7 @@ func (pgSchemaColumn *PgSchemaColumn) parquetPrimitiveTypes() (primitiveType str
 
 func (pgSchemaColumn *PgSchemaColumn) icebergPrimitiveType() string {
 	switch strings.TrimLeft(pgSchemaColumn.UdtName, "_") {
-	case "varchar", "char", "text", "interval", "jsonb", "json", "bpchar",
+	case "varchar", "char", "text", "interval", "jsonb", "json", "bpchar", "bit",
 		"point", "line", "lseg", "box", "path", "polygon", "circle",
 		"cidr", "inet", "macaddr", "macaddr8",
 		"tsvector", "pg_snapshot":

diff --git a/src/query_handler_test.go b/src/query_handler_test.go
@@ -78,6 +78,14 @@ func TestHandleQuery(t *testing.T) {
 			"description": {"count"},
 			"values":      {"2"},
 		},
+		"SELECT bit_column FROM public.test_table WHERE bit_column IS NOT NULL": {
+			"description": {"bit_column"},
+			"values":      {"1"},
+		},
+		"SELECT bit_column FROM public.test_table WHERE bit_column IS NULL": {
+			"description": {"bit_column"},
+			"values":      {""},
+		},
 		"SELECT bool_column FROM public.test_table WHERE bool_column = TRUE": {
 			"description": {"bool_column"},
 			"values":      {"true"},