Support All Data Types

[theory]

Postgres data types not currently supported or for which support is incomplete:

• bit
• bool (Add support for ClickHouse Bool #48)
• box
• bpchar
• bytea (Add support for mapping BYTEA to String #139)
• char
• cidr
• circle
• inet (Add INSERT support for UUID, INET types #134)
• int2vector
• interval
• json
• jsonb
• jsonpath
• line
• lseg
• macaddr
• macaddr8
• name
• numeric (Add support for the Decimal data type #52)
• path
• point
• polygon
• timetz
• tsquery
• tsvector
• uuid (Add INSERT support for UUID, INET types #134)
• varbit
• varchar
• xml

ClickHouse data types not currently supported or for which support is incomplete:

• BFloat16
• Date32 (Fix binary date/time conversion; add Date32 #115
• DateTime32
• Decimal (Add support for the Decimal data type #52)
• Decimal128 (Add support for the Decimal data type #52)
• Decimal256
• Decimal32 (Add support for the Decimal data type #52)
• Decimal64 (Add support for the Decimal data type #52)
• Dynamic
• Enum (converted to text)
• Enum8
• Enum16
• Int256
• IntervalDay
• IntervalHour
• IntervalMicrosecond
• IntervalMillisecond
• IntervalMinute
• IntervalMonth
• IntervalNanosecond
• IntervalQuarter
• IntervalSecond
• IntervalWeek
• IntervalYear
• IPv4 (Add INSERT support for UUID, INET types #134)
• IPv6 (Add INSERT support for UUID, INET types #134)
• JSON (Map ClickHouse JSON type to PostgreSQL JSONB #74 (http only)
• LineString
• Map
• MultiLineString
• MultiPolygon
• Nested
• Nothing
• Object
• Point
• Point
• Ring
• SimpleAggregateFunction
• Time
• Time64
• Tuple
• UInt128
• UInt256
• UUID (Add INSERT support for UUID, INET types #134)
• Variant

[WanderWang]

Hi @theory , could you please provide an estimated timeline for resolving this issue? It's currently blocking my progress, so any update would be greatly appreciated.

[theory]

Not all the types will be implemented at the same time, but will be written and released incrementally. Our initial focus is on Postgres data types mapped to ClickHouse types (working on INET now, for example), and then will turn to the remaining ClickHouse types. This work runs concurrently with our first priority, which is continuing to improve pushdown for analytics queries.