org.apache.parquet.avro.package-info Maven / Gradle / Ivy
/*
* 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.
*/
/**
*
*
* Provides classes to store Avro data in Parquet files. Avro schemas are converted to
* parquet schemas as follows. Only record schemas are converted,
* other top-level schema types are not converted and attempting to do so will result
* in an error. Avro types are converted to Parquet types using the mapping shown here:
*
*
*
*
* Avro type
* Parquet type
*
*
* null
* no type (the field is not encoded in Parquet), unless a null union
*
*
* boolean
* boolean
*
*
* int
* int32
*
*
* long
* int64
*
*
* float
* float
*
*
* double
* double
*
*
* bytes
* binary
*
*
* string
* binary (with original type UTF8)
*
*
* record
* group containing nested fields
*
*
* enum
* binary (with original type ENUM)
*
*
* array
* group (with original type LIST) containing one repeated group field
*
*
* map
* group (with original type MAP) containing one repeated group
* field (with original type MAP_KEY_VALUE) of (key, value)
*
*
* fixed
* fixed_len_byte_array
*
*
* union
* an optional type, in the case of a null union, otherwise not supported
*
*
*
*
* For Parquet files that were not written with classes from this package there is no
* Avro write schema stored in the Parquet file metadata. To read such files using
* classes from this package you must either provide an Avro read schema,
* or a default Avro schema will be derived using the following mapping.
*
*
*
* Parquet type
* Avro type
*
*
* boolean
* boolean
*
*
* int32
* int
*
*
* int64
* long
*
*
* int96
* not supported
*
*
* float
* float
*
*
* double
* double
*
*
* fixed_len_byte_array
* fixed
*
*
* binary (with no original type)
* bytes
*
*
* binary (with original type UTF8)
* string
*
*
* binary (with original type ENUM)
* string
*
*
* group (with original type LIST) containing one repeated group field
* array
*
*
* group (with original type MAP) containing one repeated group
* field (with original type MAP_KEY_VALUE) of (key, value)
* map
*
*
*
*
* Parquet fields that are optional are mapped to an Avro null union.
*
*
*
* Some conversions are lossy. Avro nulls are not represented in Parquet,
* so they are lost when converted back to Avro. Similarly, a Parquet enum does not
* store its values, so it cannot be converted back to an Avro enum,
* which is why an Avro string had to suffice. Type names for nested records, enums,
* and fixed types are lost in the conversion to Parquet.
* Avro aliases, default values, field ordering, and documentation strings are all
* dropped in the conversion to Parquet.
*
* Parquet maps can have any type for keys, but this is not true in Avro where map keys
* are assumed to be strings.
*
*/
package org.apache.parquet.avro;