@loaders.gl/schema

Table

Schemas

Batches

Table APIs

The table API is modelled after a subset of the Apache Arrow API:

Class	Arrow Counterpart	Description
`Table`	`Table`	Table
`Schema`	`Schema`	Schema
`Batch`	`RecordBatch`	Batch

Determining shape of loaded data

loaders.gl favors formats that wrap the data with a shape field so that the type of the returned data can be determined at run-time:

export type NewDataType = {
  shape: 'new-data-type';
  data: TypeOfData;
  schema?: Schema;
};

However a number of traditional return formats do not include such a wrapper.

Controlling the shape of loaded data

Loaders are encouraged to provide a shape option to allow applications to control the return format. Since different loaders offer different selection of shapes, the option is set per loader.

const tile = await load(url, MVTLoader, {mvt: {shape: 'geojson-table', ...}});
assert(tile.shape === 'geojson-table');
processTile(tile.data);

Table Category

Shape	Category	Types / Description
`table`	`Table`
`array-row-table`	`ArrayRowTable`
`object-row-table`	`ObjectRowTable`
`columnar-table`	`ColumnarTable`

Tables can be
row-oriented, i.e. organized as an array of rows
columnar, containing one array per column

Rows can contain either

an array of values, where the column name is found in the schema.
object with key-value pairs, where the key is the column name

{
  "shape": ,
  "data":
}

GIS Category

Shape	Category	Types / Description
`geojson`	`GeoJSON`	GeoJSON is a `features` array wrapped at the top level
`array-row-table`	`ArrayRowTable`
`object-row-table`	`ObjectRowTable`
`geojson-table`	`GeojsonTable`	GeoJSON table essentially contains the `features` array from the

Schemas​

Batches​

Table APIs​

Determining shape of loaded data​

Controlling the shape of loaded data​

Table Category​

GIS Category​