lamindb.base.types.Dtype

lamindb.base.types.Dtype

Data type.

String-serialized representations of common data types.

Overview

description	lamindb	pandas
categorical	"cat"	category
numerical	"num"	int | float
integer	"int"	int64 | int32 | int16 | int8 | uint | ...
float	"float"	float64 | float32 | float16 | float8 | ...
string	"str"	object
datetime	"datetime"	datetime
date	"date"	object (pandera requires an ISO-format string, convert with df["date"] = df["date"].dt.date)
dictionary	"dict"	object
path	"path"	str (pandas does not have a dedicated path type, validated as str)

Categoricals

Beyond indicating that a feature is a categorical, lamindb allows you to define the registry to which values are restricted.

For example, 'cat[ULabel]' or 'cat[bionty.CellType]' indicate that permissible values are from the ULabel or CellType registry, respectively.

You can also reference multiple registries, e.g., 'cat[ULabel|bionty.CellType]' indicates that values can be from either registry.

You can also restrict to sub-types defined in registries via the type column, e.g., 'cat[ULabel[CellMedium]]' indicates that values must be of type CellMedium within the ULabel registry.

Literal

A Dtype object in lamindb is a Literal up to further specification of "cat".

alias of Literal[‘cat’, ‘num’, ‘str’, ‘int’, ‘float’, ‘bool’, ‘date’, ‘datetime’, ‘dict’, ‘object’, ‘path’]