Delta Lake Tables API

openapi: 3.2.0
info:
  description: The Delta Sharing protocol is still very early in its development.
  version: '0.2'
  title: Delta Sharing Protocol Tables API
  contact:
    email: ''
  license:
    name: AGPL v3.0
    url: https://www.gnu.org/licenses/agpl-3.0.en.html
servers:
- url: http://localhost:8000/api/v1
  description: Local dev server (APIv1)
- url: https://sharing.delta.io/delta-sharing/
  description: Demo Delta Sharing server
tags:
- name: Tables
  description: Table query and inspection APIs
paths:
  /shares/{share}/schemas/{schema}/tables/{table}/version:
    get:
      operationId: GetTableVersion
      summary: Return the table version
      description: 'This is the API for clients to get a table version without any other

        extra information. The server usually can implement this API

        effectively. If a client caches information about a shared table

        locally, it can store the table version and use this cheap API to

        quickly check whether their cache is stale and they should re-fetch the

        data.'
      tags:
      - Tables
      parameters:
      - in: path
        name: share
        required: true
        description: Named share for finding the named schema
        schema:
          type: string
      - in: path
        name: schema
        required: true
        description: Named schema for finding the named table
        schema:
          type: string
      - in: path
        name: table
        required: true
        description: Named table
        schema:
          type: string
      - in: query
        name: startingTimestamp
        required: false
        description: Starting Timestamp
        schema:
          type: string
      responses:
        '400':
          $ref: '#/components/responses/400'
        '401':
          $ref: '#/components/responses/401'
        '403':
          $ref: '#/components/responses/403'
        '404':
          $ref: '#/components/responses/404'
        '500':
          $ref: '#/components/responses/500'
        '200':
          description: The table was found and has a version
          headers:
            Delta-Table-Version:
              description: A long value which represents the current table version
              schema:
                type: integer
  /shares/{share}/schemas/{schema}/tables/{table}/metadata:
    get:
      operationId: GetTableMetadata
      summary: Query the metadata and schema of the given table
      tags:
      - Tables
      parameters:
      - in: path
        name: share
        required: true
        description: Named share for finding the named schema
        schema:
          type: string
      - in: path
        name: schema
        required: true
        description: Named schema for finding the named table
        schema:
          type: string
      - in: path
        name: table
        required: true
        description: Named table
        schema:
          type: string
      - in: query
        name: startingTimestamp
        required: false
        description: Starting Timestamp
        schema:
          type: string
      responses:
        '400':
          $ref: '#/components/responses/400'
        '401':
          $ref: '#/components/responses/401'
        '403':
          $ref: '#/components/responses/403'
        '404':
          $ref: '#/components/responses/404'
        '500':
          $ref: '#/components/responses/500'
        '200':
          description: 'A sequence of JSON strings containing the table protocol

            and then the table metadata

            '
          headers:
            Delta-Table-Version:
              description: A long value which represents the current table version
              schema:
                type: integer
          content:
            application/x-ndjson:
              schema:
                $ref: '#/components/schemas/TableMetadataResponse'
  /shares/{share}/schemas/{schema}/tables/{table}/query:
    post:
      operationId: QueryTable
      summary: Query the table
      tags:
      - Tables
      parameters:
      - in: path
        name: share
        required: true
        description: Named share for finding the named schema
        schema:
          type: string
      - in: path
        name: schema
        required: true
        description: Named schema for finding the named table
        schema:
          type: string
      - in: path
        name: table
        required: true
        description: Named table
        schema:
          type: string
      - in: query
        name: startingTimestamp
        required: false
        description: Starting Timestamp
        schema:
          type: string
      requestBody:
        required: true
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/QueryRequest'
      responses:
        '400':
          $ref: '#/components/responses/400'
        '401':
          $ref: '#/components/responses/401'
        '403':
          $ref: '#/components/responses/403'
        '404':
          $ref: '#/components/responses/404'
        '500':
          $ref: '#/components/responses/500'
        '200':
          description: 'A sequence of JSON strings containing the table protocol, metadata,

            and then a number of lines for each file in the table.

            '
          headers:
            Delta-Table-Version:
              description: A long value which represents the current table version
              schema:
                type: integer
          content:
            application/x-ndjson:
              schema:
                $ref: '#/components/schemas/TableQueryResponse'
  /shares/{share}/schemas/{schema}/tables/{table}/changes:
    get:
      operationId: GetTableChanges
      summary: Read Change Data Feed from a Table
      description: 'This is the API for clients to read change data feed from a table.

        The API supports a start parameter and and an end parameter. The

        start/end parameter can either be a version or a timestamp. The start

        parameter must be provided. If the end parameter is not provided, the

        API will use the latest table version for it. The parameter range is

        inclusive in the query. You can specify a version as a Long or a

        timestamp as a string in the Timestamp Format.

        The change data feed represents row-level changes between versions of

        a Delta table. It records change data for UPDATE, DELETE, and MERGE operations.

        If you leverage the connectors provided by this library to read change data feed,

        it results in three metadata columns that identify the type of change event, in

        addition to the data columns:

        _change_type (type: String): There are four values: insert, update_preimage, update_postimage, delete. preimage is the value before the udpate, postimage is the value after the update.

        _commit_version (type: Long): The table version containing the change.

        _commit_timestamp (type: Long): The unix timestamp associated when the commit of the change was created, in milliseconds.'
      tags:
      - Tables
      parameters:
      - in: path
        name: share
        required: true
        description: Named share for finding the named schema
        schema:
          type: string
      - in: path
        name: schema
        required: true
        description: Named schema for finding the named table
        schema:
          type: string
      - in: path
        name: table
        required: true
        description: Named table
        schema:
          type: string
      - in: query
        name: startingTimestamp
        required: false
        description: 'The starting timestamp of the query, a string in the Timestamp Format, which will be converted to a version created greater or equal to this timestamp. '
        schema:
          type: string
      - in: query
        name: startingVersion
        required: false
        description: The starting version of the query, inclusive.
        schema:
          type: integer
      - in: query
        name: endingVersion
        required: false
        description: The ending version of the query, inclusive.
        schema:
          type: integer
      - in: query
        name: endingTimestamp
        required: false
        description: 'The starting timestamp of the query, a string in the Timestamp Format, which will be converted to a version created greater or equal to this timestamp. '
        schema:
          type: string
      - in: query
        name: includeHistoricalMetadata
        required: false
        description: If set to true, return the historical metadata if seen in the delta log. This is for the streaming client to check if the table schema is still read compatible.
        schema:
          type: boolean
      - in: query
        name: includeHistoricalProtocol
        required: false
        description: If set to true, return historical Protocol actions seen in the delta log so the streaming client can check whether the table is still read compatible (e.g. when a new reader feature is enabled mid-stream). Only effective for responses with responseformat=delta; ignored for responseformat=parquet responses.
        schema:
          type: boolean
      responses:
        '400':
          $ref: '#/components/responses/400'
        '401':
          $ref: '#/components/responses/401'
        '403':
          $ref: '#/components/responses/403'
        '404':
          $ref: '#/components/responses/404'
        '500':
          $ref: '#/components/responses/500'
        '200':
          description: This is the API for clients to read change data feed from a table.
          content:
            application/x-ndjson:
              schema:
                $ref: '#/components/schemas/ChangeDataFeedResponse'
components:
  schemas:
    QueryRequest:
      type: object
      properties:
        predicateHints:
          type: array
          description: 'a list of SQL boolean expressions using a restricted subset of SQL, in a JSON array.

            When it’s present, the server will use the provided predicates as a hint to apply the

            SQL predicates on the returned files. Filtering files based on the SQL predicates

            is BEST EFFORT. The server may return files that don’t satisfy the predicates.

            If the server fails to parse one of the SQL predicates, or fails to evaluate it, the

            server may skip it. Predicate expressions are conjunctive (AND-ed together).

            When it’s absent, the server will return all of files in the table.

            This will be deprecated once all the client and server implementation move to using jsonPredicateHints below.

            '
          example:
          - col = 123
          items:
            type: string
        jsonPredicateHints:
          type: object
          description: 'query predicates on partition columns specified using a structured JSON format.

            When it’s present, the server will try to use the predicates to filter table''s

            files, which could boost query performance.

            As with predicateHints, this filtering is BEST EFFORT. The server may return files

            that don’t satisfy the predicates.

            If the server encounters any errors during predicate processing (for example, invalid

            syntax or non existing columns), it will skip filtering and return all the files.

            When it’s absent, the server will return all the files in the table.

            '
          properties:
            op:
              $ref: '#/components/schemas/Ops'
            children:
              type: string
            name:
              type: string
            value:
              type: string
            valueType:
              type: string
        limitHint:
          type: integer
          example: 1000
          description: 'It’s a hint from the client to tell the server how many rows the

            client plans to read. The server can use this hint to return only

            some files by using the stats in the Delta transaction logs. For

            example, when running SELECT * FROM table LIMIT 1000, the client

            can send limit=1000 to the server

            '
        version:
          type: integer
          description: 'an optional version number. If set, will return files as of the

            specified version of the table. This is only supported on tables

            with history sharing enabled.

            '
          example: 1005
        timestamp:
          type: string
          example: yyyy-[m]m-[d]d hh:mm:ss[.f...]
          description: 'an optional timestamp string in the Timestamp Format,. If set, will

            return files as of the table version corresponding to the specified

            timestamp. This is only supported on tables with history sharing enabled.

            '
        startingVersion:
          type: integer
          example: 1000
          description: 'an optional version number. If set, will return all data change files

            since startingVersion, inclusive, including historical metadata if seen

            in the delta log.

            '
        endingVersion:
          type: integer
          example: 1000
          description: 'an optional version number, only used if startingVersion is set. If set,

            the server can use it as a hint to avoid returning data change files after

            endingVersion. This is not enforcement. Hence, when sending the endingVersion

            parameter, the client should still handle the case that it may receive files

            after endingVersion.

            '
        includeHistoricalProtocol:
          type: boolean
          example: true
          description: 'an optional boolean, only used if startingVersion is set. If true, the

            server inlines historical Protocol actions seen in the delta log (for

            versions strictly after startingVersion) so the streaming client can check

            whether the table is still read compatible (e.g. when a new reader feature

            is enabled mid-stream). Only effective for responses with

            responseformat=delta; ignored for responseformat=parquet responses.

            '
    TableMetadataResponse:
      type: string
      example: '{"protocol":{"minReaderVersion":1}}

        {"metaData":{"id":"f8d5c169-3d01-4ca3-ad9e-7dc3355aedb2","format":{"provider":"parquet"},"schemaString":"{\"type\":\"struct\",\"fields\":[{\"name\":\"eventTime\",\"type\":\"timestamp\",\"nullable\":true,\"metadata\":{}},{\"name\":\"date\",\"type\":\"date\",\"nullable\":true,\"metadata\":{}}]}","partitionColumns":["date"]}}

        '
    Ops:
      type: string
      enum:
      - column
      - literal
      - isNull
      - equal
      - lessThan
      - lessThanOrEqual
      - greaterThan
      - greaterThanOrEqual
      - and
      - or
      - not
      description: "Op \t                Op Type \tDescription\ncolumn \t            Leaf \tRepresents a column. A column op has two fields: name and valueType. The column's value will be cast to the specified valueType during comparisons.\nliteral \t        Leaf    Represents a literal or fixed value. A literal op has two fields: value and valueType. The literal's value will be cast to the specified valueType during comparisons.\nisNull \t            Unary \tRepresents a null check on a column op. This op should have only one child, the column op.\nequal \t            Binary \tRepresents an equality (\"=\") check. This op should have two children, and both should be leaf ops.\nlessThan \t        Binary \tRepresents a less than (\"<\") check. This op should have two children, and both should be leaf ops.\nlessThanOrEqual\t    Binary \tRepresents a less than or equal (\"<=\") check. This op should have two children, and both should be leaf ops.\ngreaterThan \t    Binary \tRepresents a greater than (\">\") check. This op should have two children, and both should be leaf ops.\ngreaterThanOrEqual \tBinary \tRepresents a greater than (\">=\") check. This op should have two children, and both should be leaf ops.\nand \t            Nary \tRepresents a logical and operation amongst its children. This op should have at least two children.\nor \t                Nary \tRepresents a logical or operation amongst its children. This op should have at least two children.\nnot \t            Unary \tRepresents a logical not check. This op should have once child.\nThe supported value types:\n  ValueType \tDescription\n  \"bool\" \tRepresents an Boolean type.\n  \"int\" \tRepresents an Integer type.\n  \"long\" \tRepresents a Long type.\n  \"string\" \tRepresents a String type.\n  \"date\" \tRepresents a Date type in \"yyyy-mm-dd\" format.\n"
    ChangeDataFeedResponse:
      type: string
      example: '{"protocol": {"minReaderVersion": 1}}

        {"metaData": {"id": "f8d5c169-3d01-4ca3-ad9e-7dc3355aedb2","format": {"provider": "parquet"},"schemaString": "{\"type\":\"struct\",\"fields\":[{\"name\":\"eventTime\",\"type\":\"timestamp\",\"nullable\":true,\"metadata\":{}},{\"name\":\"date\",\"type\":\"date\",\"nullable\":true,\"metadata\":{}}]}","partitionColumns": ["date"],"configuration": {"enableChangeDataFeed": "true"}}}{"add": {"url": "https://<s3-bucket-name>.s3.us-west-2.amazonaws.com/delta-exchange-test/table_cdf/date%3D2021-04-28/part-00000-8b0086f2-7b27-4935-ac5a-8ed6215a6640.c000.snappy.parquet?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20210501T010516Z&X-Amz-SignedHeaders=host&X-Amz-Expires=900&X-Amz-Credential=REDACTED_AWS_ACCESS_KEY_ID%2F20210501%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Signature=97b6762cfd8e4d7e94b9d707eff3faf266974f6e7030095c1d4a66350cfd892e","id": "8b0086f2-7b27-4935-ac5a-8ed6215a6640","partitionValues": {"date": "2021-04-28"},"size":573,"stats": "{\"numRecords\":1,\"minValues\":{\"eventTime\":\"2021-04-28T23:33:57.955Z\"},\"maxValues\":{\"eventTime\":\"2021-04-28T23:33:57.955Z\"},\"nullCount\":{\"eventTime\":0}}","timestamp": 1652140000000,"version": 0}}

        {"cdf": {"url": "https://<s3-bucket-name>.s3.us-west-2.amazonaws.com/delta-exchange-test/table_cdf/_change_data/date%3D2021-04-28/part-00000-591723a8-6a27-4240-a90e-57426f4736d2.c000.snappy.parquet?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20210501T010516Z&X-Amz-SignedHeaders=host&X-Amz-Expires=899&X-Amz-Credential=REDACTED_AWS_ACCESS_KEY_ID%2F20210501%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Signature=0f7acecba5df7652457164533a58004936586186c56425d9d53c52db574f6b62","id": "591723a8-6a27-4240-a90e-57426f4736d2","partitionValues": {"date": "2021-04-28"},"size": 689,"timestamp": 1652141000000,"version": 1}}

        {"remove": {"url": "https://<s3-bucket-name>.s3.us-west-2.amazonaws.com/delta-exchange-test/table_cdf/date%3D2021-04-28/part-00000-8b0086f2-7b27-4935-ac5a-8ed6215a6640.c000.snappy.parquet?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20210501T010516Z&X-Amz-SignedHeaders=host&X-Amz-Expires=900&X-Amz-Credential=REDACTED_AWS_ACCESS_KEY_ID%2F20210501%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Signature=97b6762cfd8e4d7e94b9d707eff3faf266974f6e7030095c1d4a66350cfd892e","id": "8b0086f2-7b27-4935-ac5a-8ed6215a6640","partitionValues": {"date": "2021-04-28"},"size": 573,"timestamp": 1652142000000,"version": 2}}

        '
    TableQueryResponse:
      type: string
      example: '{"protocol":{"minReaderVersion":1}}

        {"metaData":{"id":"f8d5c169-3d01-4ca3-ad9e-7dc3355aedb2","format":{"provider":"parquet"},"schemaString":"{\"type\":\"struct\",\"fields\":[{\"name\":\"eventTime\",\"type\":\"timestamp\",\"nullable\":true,\"metadata\":{}},{\"name\":\"date\",\"type\":\"date\",\"nullable\":true,\"metadata\":{}}]}","partitionColumns":["date"]}}

        {"file":{"url":"https://<s3-bucket-name>.s3.us-west-2.amazonaws.com/delta-exchange-test/table2/date%3D2021-04-28/part-00000-8b0086f2-7b27-4935-ac5a-8ed6215a6640.c000.snappy.parquet?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20210501T010516Z&X-Amz-SignedHeaders=host&X-Amz-Expires=900&X-Amz-Credential=REDACTED_AWS_ACCESS_KEY_ID%2F20210501%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Signature=97b6762cfd8e4d7e94b9d707eff3faf266974f6e7030095c1d4a66350cfd892e","id":"8b0086f2-7b27-4935-ac5a-8ed6215a6640","partitionValues":{"date":"2021-04-28"},"size":573,"stats":"{\"numRecords\":1,\"minValues\":{\"eventTime\":\"2021-04-28T23:33:57.955Z\"},\"maxValues\":{\"eventTime\":\"2021-04-28T23:33:57.955Z\"},\"nullCount\":{\"eventTime\":0}}"}}

        {"file":{"url":"https://<s3-bucket-name>.s3.us-west-2.amazonaws.com/delta-exchange-test/table2/date%3D2021-04-28/part-00000-591723a8-6a27-4240-a90e-57426f4736d2.c000.snappy.parquet?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Date=20210501T010516Z&X-Amz-SignedHeaders=host&X-Amz-Expires=899&X-Amz-Credential=REDACTED_AWS_ACCESS_KEY_ID%2F20210501%2Fus-west-2%2Fs3%2Faws4_request&X-Amz-Signature=0f7acecba5df7652457164533a58004936586186c56425d9d53c52db574f6b62","id":"591723a8-6a27-4240-a90e-57426f4736d2","partitionValues":{"date":"2021-04-28"},"size":573,"stats":"{\"numRecords\":1,\"minValues\":{\"eventTime\":\"2021-04-28T23:33:48.719Z\"},\"maxValues\":{\"eventTime\":\"2021-04-28T23:33:48.719Z\"},\"nullCount\":{\"eventTime\":0}}"}}

        '
  securitySchemes:
    BearerAuth:
      type: http
      scheme: bearer