Schema generation

This page documents the API for generating table schemas that are compatible with both upsert and insert.

builder

builder(schema, *, key=None)

Generate an empty schema for a keyed or unkeyed q table.

Parameters:

Name	Type	Description	Default
schema	Dict	The definition of the schema to be created mapping a 'str' to a pykx.* type object. Each pykx.* value must be one of the types defined in pykx.schema._ktype_to_conversion.	required
key	Optional[Union[str, List[str]]]	The column name(s) in schema to be treated as primary keys.	None

Returns:

Type	Description
k.K	A pykx.Table or pykx.KeyedTable matching the provided schema with zero rows.

Examples:

Create a simple pykx.Table with four columns of different types

>>> import pykx as kx
>>> qtab = kx.schema.builder({
        'col1' : kx.GUIDAtom,
        'col2': kx.TimeAtom,
        'col3': kx.BooleanAtom,
        'col4': kx.FloatAtom
        })
>>> qtab
pykx.Table(pykx.q('
col1 col2 col3 col4
-------------------
'))
>>> kx.q.meta(qtab)
pykx.KeyedTable(pykx.q('
c   | t f a
----| -----
col1| g
col2| t
col3| b
col4| f
'))

Create a pykx.KeyedTable with a single primary key.

>>> import pykx as kx
>>> qtab = kx.schema.builder({
        'col1': kx.TimestampAtom,
        'col2': kx.FloatAtom,
        'col3': kx.IntAtom},
        key = 'col1'
        )
>>> qtab
pykx.KeyedTable(pykx.q('
col1| col2 col3
----| ---------
'))
>>> kx.q.meta(qtab)
pykx.KeyedTable(pykx.q('
c   | t f a
----| -----
col1| p
col2| f
col3| i
'))

Create a pykx.KeyedTable with multiple primary keys.

>>> import pykx as kx
>>> qtab = kx.schema.builder({
        'col1': kx.TimestampAtom,
        'col2': kx.SymbolAtom,
        'col3': kx.IntAtom,
        'col4': kx.List},
        key = ['col1', 'col2']
        )
>>> qtab
pykx.KeyedTable(pykx.q('
col1 col2| col3 col4
---------| ---------
'))
>>> kx.q.meta(qtab)
pykx.KeyedTable(pykx.q('
c   | t f a
----| -----
col1| p
col2| s
col3| i
col4|
'))