amoeba commented on code in PR #41593: URL: https://github.com/apache/arrow/pull/41593#discussion_r1602358811
########## docs/source/format/FormatIntro.rst: ########## @@ -0,0 +1,511 @@ +.. Licensed to the Apache Software Foundation (ASF) under one +.. or more contributor license agreements. See the NOTICE file +.. distributed with this work for additional information +.. regarding copyright ownership. The ASF licenses this file +.. to you under the Apache License, Version 2.0 (the +.. "License"); you may not use this file except in compliance +.. with the License. You may obtain a copy of the License at + +.. http://www.apache.org/licenses/LICENSE-2.0 + +.. Unless required by applicable law or agreed to in writing, +.. software distributed under the License is distributed on an +.. "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +.. KIND, either express or implied. See the License for the +.. specific language governing permissions and limitations +.. under the License. + +***************************************** +Introduction to the Arrow Columnar Format +***************************************** + +Apache Arrow was born with the idea to define a set of standards for +data representation and interchange between languages and systems to +avoid costs of data serialization/deserialization and in order to +avoid reinventing the wheel on each of those systems and languages. + +Each system / language requires their own format definitions, implementation +of common algorithms, etcetera. In our heterogeneous environments we +often have to move data from one system/language to accommodate our +workflows that meant copy&convert the data between them, which is +quite costly. + +Apart from this initial vision, Arrow has grown to also develop a +multi-language collection of libraries for solving systems problems +related to in-memory analytical data processing. This includes such +topics as: + +* Zero-copy shared memory and RPC-based data movement +* Reading and writing file formats (like CSV, Apache ORC, and Apache Parquet) +* In-memory analytics and query processing + +Arrow Columnar Format +===================== + +.. figure:: ./images/columnar-diagram_1.svg + :scale: 70% + :alt: Diagram with tabular data of 4 rows and columns. + +Data can be represented in memory using a row based format or a column +based format. The row based format stores data by row meaning the rows +are adjacent in the computer memory: + +.. figure:: ./images/columnar-diagram_2.svg + :alt: Tabular data being structured row by row in computer memory. + +In a columnar format, on the other hand, the data is organised by column +instead of by row making analytical operations like filtering, grouping, +aggregations and others much more efficient. CPU can maintain memory locality +and require less memory jumps to process the data. By keeping the data contiguous +in memory it also enables vectorization of the computations. Most modern +CPUs have single instructions, multiple data (SIMD) enabling parallel +processing and execution of instructions on vector data in single CPU +instructions. + +.. figure:: ./images/columnar-diagram_3.svg + :alt: Tabular data being structured column by column in computer memory. + +Compression is another element where columnar format representation can +take high advantage. Data similarity allows for better compression +techniques and algorithms. Having the same data types locality allows +to have better compression ratios. + +Primitive layouts +================= + +Fixed Size Primitive Layout +--------------------------- + +A primitive column represents an array of values where each value +has the same physical size measured in bytes. Data types that share the +same fixed size primitive layout are for example signed and unsigned +integer types, floating point numbers, boolean, decimal and temporal +types. + +Support for null values +----------------------- + +Arrow supports missing values or "nulls" for all data types: any value +in an array may be semantically null, whether primitive or nested type. + +In Arrow, a dedicated buffer, known as the validity (or "null") bitmap, +is used alongside the data indicating whether each value in the array is +null or not. You can think of it as vector of 0 and 1 values, where a 1 +means that the value is not-null ("valid"), while a 0 indicates the value +is null. + +This validity bitmap is optional, i.e. if there are no missing values in +the array the buffer does not need to be allocated (as in the example +column 1 in the diagram below). + +.. figure:: ./images/primitive-diagram.svg + :alt: Diagram is showing the difference between the primitive data + type presented in a Table and the data actually stored in + computer memory. + + Physical layout diagram for primitive data types. + +.. note:: + Boolean data type is represented with a primitive layout where the + values are encoded in bits instead of bytes. That means the physical + layout includes a values bitmap buffer and possibly a validity bitmap + buffer. + + .. figure:: ./images/bool-diagram.svg + :alt: Diagram is showing the difference between the boolean data + type presented in a Table and the data actually stored in + computer memory. + + Physical layout diagram for boolean data type. + +.. note:: + Arrow also has a concept of Null type where all values are null. In + this case no memory buffers are allocated. + +Variable length binary and string +--------------------------------- + +The bytes of a binary or string column are stored together consecutively +in a single buffer or region of memory. To know where each element of the +column starts and ends the physical layout also includes integer offsets. +The length of the offset buffer is one more than the length of the values +buffer as the last two elements define the start and the end of the last +element in the binary/string column. + +Binary and string types share the same physical layout. The one difference +between them is that the string type is utf-8 binary and will produce an +invalid result if the bytes are not valid utf-8. + +The difference between binary/string and large binary/string is in the offset +type. In the first case that is int32 and in the second it is int64. + +The limitation of types using 32 bit offsets is that they have a max size of +2GB per array. One can still use the non-large variants for bigger data, but +then multiple chunks are needed. + +.. figure:: ./images/var-string-diagram.svg + :alt: Diagram is showing the difference between the variable length + string data type presented in a Table and the data actually + stored in computer memory. + + Physical layout diagram for variable length string data types. + +Variable length binary and string view +-------------------------------------- + +This layout is adapted from TU Munich's UmbraDB and is similar to the string +layout used in DuckDB and Velox (and sometimes also called "German style strings"). + +The main differences to classical binary and string types is the views buffer. +It includes the length of the string, and then either contains the characters +inline (for small strings) or only the first 4 bytes of the string and point to +potentially several data buffers. It also supports binary and strings to be written +out of order. + +These properties are important for efficient string processing. The prefix +enables a profitable fast path for string comparisons, which are frequently +determined within the first four bytes. Selecting elements is a simple "take" +operation on the fixed-width views buffer and does not need to rewrite the +values buffers. + +.. figure:: ./images/var-string-view-diagram.svg + :alt: Diagram is showing the difference between the variable length + string view data type presented in a Table and the dataactually + stored in computer memory. + + Physical layout diagram for variable length string view data type. + +Nested layouts +============== + +Nested types introduce the concept of parent and child arrays. They express +relationships between physical value arrays in a nested type structure. + +Nested types depend on one or more other child data types. For instance, List +is a nested type (parent) that has one child (the data types of the values in +the list). + +List +---- + +The list type enables values of the same type being stacked together in a +sequence of values in each column slot. The layout is similar to binary or +string type as it has offsets buffer to define where the sequence of values +starts and ends with all the values of the column being stored consecutively +in a values child array. + +The offsets in the list type are int32 while in the large list the offsets +are int64. + +.. figure:: ./images/var-list-diagram.svg + :alt: Diagram is showing the difference between the variable size + list data type presented in a Table and the dataactually + stored in computer memory. + + Physical layout diagram for variable size list data type. + +Fixed size list +--------------- + +Fixed size list is a special case of variable-size list where each column slot +contains a fixed size sequence meaning all lists are the same size and so the +offset buffer is no longer needed. + +.. figure:: ./images/fixed-list-diagram.svg + :alt: Diagram is showing the difference between the fixed size list data + type presented in a Table and the dataactually stored in computer + memory. + + Physical layout diagram for fixed size list data type. + +List and large list view +------------------------ + +List view type allows arrays to specify out-of-order offsets. + +.. figure:: ./images/var-list-view-diagram.svg + :alt: Diagram is showing the difference between the variable size list view + data type presented in a Table and the dataactually stored in + computer memory. + + Physical layout diagram for variable size list view data type. + +Struct +------ + +A struct is a nested type parameterized by an ordered sequence of types. + +* There is one child array for each field +* Child arrays are independent and need not be adjacent to each other in + memory (only need to have the same length) + +One can think of an individual struct field as a key-value pair where the +key is the field name and the child array its values. The field (key) is +saved in the schema and the values of a specific field (key) are saved in +the child array. + +.. figure:: ./images/struct-diagram.svg + :alt: Diagram is showing the difference between the struct data type + presented in a Table and the dataactually stored in computer + memory. + + Physical layout diagram for struct data type. + +Map +--- + +Map type represents nested data where each value is a variable number of +key-value pairs. Its physical representation is the same as a list of ``{key, value}`` +structs. + +The difference between a struct and a map type is that a struct holds the key +in the schema, therefore need to be a string, and the values in the child arrays, +one for each field. There can be multiple keys and therefore multiple child arrays. +The map, on the other hand, has one child array holding all the different keys (that +thus all need to be of the same type but not necessarily strings) and a second +child array holding all the values, those values need to be of the same type (which +doesn't have to match the one on the keys). + +Also, the map stores the struct in a list and needs an offset as the list is +variable shape. + +.. figure:: ./images/map-diagram.svg + :alt: Diagram is showing the difference between the map data type + presented in a Table and the dataactually stored in computer + memory. + + Physical layout diagram for map data type. + +Union +----- + +The union is a nested type where each slot in the union has a value with a type chosen +from a subset of possible Arrow data types. That means that a union array represents a +mixed-type array. Unlike other data types, unions do not have their own validity bitmap +and the nullness is determined by the child arrays. + +Arrow defines two distinct union types, “dense” and “sparse”. + +Dense Union +^^^^^^^^^^^ + +Dense Union has one child array for each type present in the mixed-type array and + +* **Types buffer:** holds type id for each slot of the array. Type id corresponds + to the number of the child array. +* **Offsets buffer:** holds relative offset into the respective child array for each + array slot. + +.. figure:: ./images/dense-union-diagram.svg + :alt: Diagram is showing the difference between the dense union data type + presented in a Table and the dataactually stored in computer + memory. + + Physical layout diagram for dense union data type. + +Sparse union +^^^^^^^^^^^^ + +A sparse union has the same structure as a dense union, with the omission of the offsets +array. In this case, the child arrays are each equal in length to the length of the union. + + +.. figure:: ./images/sparse-union-diagram.svg + :alt: Diagram is showing the difference between the sparse union data type + presented in a Table and the dataactually stored in computer + memory. + + Physical layout diagram for sparse union data type. + +Dictionary Encoded Layout +========================= + +Dictionary encoding can be effective when you have data with many repeated values. +The values are represented by integers referencing a dictionary usually consisting of +unique values. + +.. figure:: ./images/dictionary-diagram.svg + :alt: Diagram is showing the difference between the dictionary data type + presented in a Table and the dataactually stored in computer + memory. + + Physical layout diagram for dictionary data type. + +Run-End Encoded Layout +====================== + +Run-end encoding is well-suited for representing data containing sequences of the +same value. These sequences are called runs. Run-end encoded array has no buffers +by itself, but has two child arrays: + +* **Run ends array:** holds the index in the array where each run ends. +* **Values array:** the actual values without repetitions. + +.. figure:: ./images/ree-diagram.svg + :alt: Diagram is showing the difference between the run-end encoded data + type presented in a Table and the dataactually stored in computer + memory. + + Physical layout diagram for run-end encoded data type. + +.. link to All types overview https://github.com/apache/arrow/issues/14752 + +Extension Types +=============== + +In case the system or application needs to extend standard Arrow data types with +custom semantics this is enabled by defining extension types or user-defined types. + +For example: + +* Universally unique identifier (uuid) can be represented as a FixedSizeBinary type +* Trading time can be represented as a Timestamp with metadata indicating the market + trading calendar + +Extension types can be defined by annotating any of the built-in Arrow logical types +(the “storage type”) with a custom type name and optional serialized representation +(``'ARROW:extension:name'`` and ``'ARROW:extension:metadata'`` keys in the Field +metadata structure). + +.. seealso:: + The :ref:`format_metadata_extension_types` documentation. + +Canonical Extension Types +------------------------- + +It is beneficial to share the definitions of well-known extension types so as to +improve interoperability between different systems integrating Arrow columnar data. +For this reason canonical extension types are defined in Arrow itself. + +Examples: + +* Fixed and variable shape tensor + + - :ref:`fixed_shape_tensor_extension` + - :ref:`variable_shape_tensor_extension` + +.. seealso:: + The :ref:`format_canonical_extensions` documentation. + +Community Extension Types +------------------------- +These are Arrow extension types that have been established as standards within specific domain areas. + +Example: + +* GeoArrow - collection of Arrow extension types for representing vector geometries + https://github.com/geoarrow/geoarrow + +Overview of Arrow terminology +============================= + +Buffer +------ +A contiguous region of memory with a given length. Buffers are used to store data for arrays. + +Array +----- +A contiguous, one-dimensional sequence of values with known length where all values have the +same type. An array consists of zero or more buffers. + +Chunked Array +------------- +A discontiguous, one-dimensional sequence of values with known length where all values have +the same type. Consists of zero or more arrays, the “chunks”. + +.. note:: + Chunked array is a concept specific to certain implementations such as Arrow C++ and PyArrow. + +RecordBatch +----------- +A contiguous, two-dimensional data structure which consist of ordered collection of arrays +of the same length. + +Schema +------ +A collection of fields with optional metadata that determines all the data types of an object +like a record batch or table. Review Comment: ```suggestion like a RecordBatch or Table. ``` -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: github-unsubscr...@arrow.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
