• This will become a table of contents (this text will be scraped). {:toc}

Since Spark 2.4 release, Spark SQL provides built-in support for reading and writing Apache Avro data.

Deploying

The spark-avro module is external and not included in spark-submit or spark-shell by default.

As with any Spark applications, spark-submit is used to launch your application. spark-avro_{{site.SCALA_BINARY_VERSION}} and its dependencies can be directly added to spark-submit using --packages, such as,

For experimenting on spark-shell, you can also use --packages to add org.apache.spark:spark-avro_{{site.SCALA_BINARY_VERSION}} and its dependencies directly,

See Application Submission Guide for more details about submitting applications with external dependencies.

Load and Save Functions

Since spark-avro module is external, there is no .avro API in DataFrameReader or DataFrameWriter.

To load/save data in Avro format, you need to specify the data source option format as avro(or org.apache.spark.sql.avro).

<div class="codetabs"> <div data-lang="python" markdown="1"> {% highlight python %}

df = spark.read.format("avro").load("examples/src/main/resources/users.avro") df.select("name", "favorite_color").write.format("avro").save("namesAndFavColors.avro")

{% endhighlight %}

</div> <div data-lang="scala" markdown="1"> {% highlight scala %}

val usersDF = spark.read.format("avro").load("examples/src/main/resources/users.avro") usersDF.select("name", "favorite_color").write.format("avro").save("namesAndFavColors.avro")

{% endhighlight %}

</div> <div data-lang="java" markdown="1"> {% highlight java %}

Dataset<Row> usersDF = spark.read().format("avro").load("examples/src/main/resources/users.avro"); usersDF.select("name", "favorite_color").write().format("avro").save("namesAndFavColors.avro");

{% endhighlight %}

</div> <div data-lang="r" markdown="1"> {% highlight r %}

df <- read.df("examples/src/main/resources/users.avro", "avro") write.df(select(df, "name", "favorite_color"), "namesAndFavColors.avro", "avro")

{% endhighlight %}

</div> </div>

to_avro() and from_avro()

The Avro package provides function to_avro to encode a column as binary in Avro format, and from_avro() to decode Avro binary data into a column. Both functions transform one column to another column, and the input/output SQL data type can be a complex type or a primitive type.

Using Avro record as columns is useful when reading from or writing to a streaming source like Kafka. Each Kafka key-value record will be augmented with some metadata, such as the ingestion timestamp into Kafka, the offset in Kafka, etc.

• If the "value" field that contains your data is in Avro, you could use from_avro() to extract your data, enrich it, clean it, and then push it downstream to Kafka again or write it out to a file.
• to_avro() can be used to turn structs into Avro records. This method is particularly useful when you would like to re-encode multiple columns into a single one when writing data out to Kafka.

<div class="codetabs"> <div data-lang="python" markdown="1"> {% highlight python %} from pyspark.sql.avro.functions import from_avro, to_avro

from_avro requires Avro schema in JSON string format.

jsonFormatSchema = open("examples/src/main/resources/user.avsc", "r").read()

df = spark
.readStream
.format("kafka")
.option("kafka.bootstrap.servers", "host1:port1,host2:port2")
.option("subscribe", "topic1")
.load()

1. Decode the Avro data into a struct;

2. Filter by column favorite_color;

3. Encode the column name in Avro format.

output = df
.select(from_avro("value", jsonFormatSchema).alias("user"))
.where('user.favorite_color == "red"')
.select(to_avro("user.name").alias("value"))

query = output
.writeStream
.format("kafka")
.option("kafka.bootstrap.servers", "host1:port1,host2:port2")
.option("topic", "topic2")
.start()

{% endhighlight %}

</div> <div data-lang="scala" markdown="1"> {% highlight scala %} import org.apache.spark.sql.avro.functions._

// from_avro requires Avro schema in JSON string format. val jsonFormatSchema = new String(Files.readAllBytes(Paths.get("./examples/src/main/resources/user.avsc")))

val df = spark .readStream .format("kafka") .option("kafka.bootstrap.servers", "host1:port1,host2:port2") .option("subscribe", "topic1") .load()

// 1. Decode the Avro data into a struct; // 2. Filter by column favorite_color; // 3. Encode the column name in Avro format. val output = df .select(from_avro($"value", jsonFormatSchema) as $"user") .where("user.favorite_color == "red"") .select(to_avro($"user.name") as $"value")

val query = output .writeStream .format("kafka") .option("kafka.bootstrap.servers", "host1:port1,host2:port2") .option("topic", "topic2") .start()

{% endhighlight %}

</div> <div data-lang="java" markdown="1"> {% highlight java %} import static org.apache.spark.sql.functions.col; import static org.apache.spark.sql.avro.functions.*;

// from_avro requires Avro schema in JSON string format. String jsonFormatSchema = new String(Files.readAllBytes(Paths.get("./examples/src/main/resources/user.avsc")));

Dataset<Row> df = spark .readStream() .format("kafka") .option("kafka.bootstrap.servers", "host1:port1,host2:port2") .option("subscribe", "topic1") .load();

// 1. Decode the Avro data into a struct; // 2. Filter by column favorite_color; // 3. Encode the column name in Avro format. Dataset<Row> output = df .select(from_avro(col("value"), jsonFormatSchema).as("user")) .where("user.favorite_color == "red"") .select(to_avro(col("user.name")).as("value"));

StreamingQuery query = output .writeStream() .format("kafka") .option("kafka.bootstrap.servers", "host1:port1,host2:port2") .option("topic", "topic2") .start();

{% endhighlight %}

</div> <div data-lang="r" markdown="1"> {% highlight r %}

from_avro requires Avro schema in JSON string format.

jsonFormatSchema <- paste0(readLines("examples/src/main/resources/user.avsc"), collapse=" ")

df <- read.stream( "kafka", kafka.bootstrap.servers = "host1:port1,host2:port2", subscribe = "topic1" )

1. Decode the Avro data into a struct;

2. Filter by column favorite_color;

3. Encode the column name in Avro format.

output <- select( filter( select(df, alias(from_avro("value", jsonFormatSchema), "user")), column("user.favorite_color") == "red" ), alias(to_avro("user.name"), "value") )

write.stream( output, "kafka", kafka.bootstrap.servers = "host1:port1,host2:port2", topic = "topic2" ) {% endhighlight %}

</div> <div data-lang="sql" markdown="1"> {% highlight sql %} CREATE TABLE t AS SELECT NAMED_STRUCT('u', NAMED_STRUCT('member0', member0, 'member1', member1)) AS s FROM VALUES (1, NULL), (NULL, 'a') tab(member0, member1); DECLARE avro_schema STRING; SET VARIABLE avro_schema = '{ "type": "record", "name": "struct", "fields": [{ "name": "u", "type": ["int","string"] }] }';

SELECT TO_AVRO(s, avro_schema) AS RESULT FROM t;

SELECT FROM_AVRO(result, avro_schema, MAP()).u FROM ( SELECT TO_AVRO(s, avro_schema) AS RESULT FROM t);

DROP TEMPORARY VARIABLE avro_schema; DROP TABLE t; {% endhighlight %}

</div> </div>

Data Source Option

Data source options of Avro can be set via:

• the .option method on DataFrameReader or DataFrameWriter.
• the options parameter in function from_avro.

<table class="spark-config"> <thead><tr><th>Property Name</th><th>Default</th><th>Meaning</th><th>Scope</th><th>Since Version</th></tr></thead> <tr> <td><code>avroSchema</code></td> <td>None</td> <td>Optional schema provided by a user in JSON format. <ul> <li> When reading Avro files or calling function <code>from_avro</code>, this option can be set to an evolved schema, which is compatible but different with the actual Avro schema. The deserialization schema will be consistent with the evolved schema. For example, if we set an evolved schema containing one additional column with a default value, the reading result in Spark will contain the new column too. Note that when using this option with <code>from_avro</code>, you still need to pass the actual Avro schema as a parameter to the function. </li> <li> When writing Avro, this option can be set if the expected output Avro schema doesn't match the schema converted by Spark. For example, the expected schema of one column is of "enum" type, instead of "string" type in the default converted schema. </li> </ul> </td> <td> read, write and function <code>from_avro</code></td> <td>2.4.0</td> </tr> <tr> <td><code>recordName</code></td> <td>topLevelRecord</td> <td>Top level record name in write result, which is required in Avro spec.</td> <td>write</td> <td>2.4.0</td> </tr> <tr> <td><code>recordNamespace</code></td> <td>""</td> <td>Record namespace in write result.</td> <td>write</td> <td>2.4.0</td> </tr> <tr> <td><code>ignoreExtension</code></td> <td>true</td> <td>The option controls ignoring of files without <code>.avro</code> extensions in read. If the option is enabled, all files (with and without <code>.avro</code> extension) are loaded. The option has been deprecated, and it will be removed in the future releases. Please use the general data source option <a href="./sql-data-sources-generic-options.html#path-glob-filter">pathGlobFilter</a> for filtering file names.</td> <td>read</td> <td>2.4.0</td> </tr> <tr> <td><code>compression</code></td> <td>snappy</td> <td>The <code>compression</code> option allows to specify a compression codec used in write.

Currently supported codecs are <code>uncompressed</code>, <code>snappy</code>, <code>deflate</code>, <code>bzip2</code>, <code>xz</code> and <code>zstandard</code>. If the option is not set, the configuration <code>spark.sql.avro.compression.codec</code> config is taken into account.</td> <td>write</td> <td>2.4.0</td>

</tr> <tr> <td><code>mode</code></td> <td>FAILFAST</td> <td>The <code>mode</code> option allows to specify parse mode for function <code>from_avro</code>. </tr> <tr> <td><code>datetimeRebaseMode</code></td> <td>(value of <code>spark.sql.avro.datetimeRebaseModeInRead</code> configuration)</td> <td>The <code>datetimeRebaseMode</code> option allows to specify the rebasing mode for the values of the <code>date</code>, <code>timestamp-micros</code>, <code>timestamp-millis</code> logical types from the Julian to Proleptic Gregorian calendar. </tr> <tr> <td><code>positionalFieldMatching</code></td> <td>false</td> <td>This can be used in tandem with the `avroSchema` option to adjust the behavior for matching the fields in the provided Avro schema with those in the SQL schema. By default, the matching will be performed using field names, ignoring their positions. If this option is set to "true", the matching will be based on the position of the fields.</td> <td>read and write</td> <td>3.2.0</td> </tr> <tr> <td><code>enableStableIdentifiersForUnionType</code></td> <td>false</td> <td>If it is set to true, Avro schema is deserialized into Spark SQL schema, and the Avro Union type is transformed into a structure where the field names remain consistent with their respective types. The resulting field names are converted to lowercase, e.g. member_int or member_string. If two user-defined type names or a user-defined type name and a built-in type name are identical regardless of case, an exception will be raised. However, in other cases, the field names can be uniquely identified.</td> <td>read</td> <td>3.5.0</td> </tr> <tr> <td><code>stableIdentifierPrefixForUnionType</code></td> <td>member_</td> <td>When `enableStableIdentifiersForUnionType` is enabled, the option allows to configure the prefix for fields of Avro Union type.</td> <td>read</td> <td>4.0.0</td> </tr> <tr> <td><code>recursiveFieldMaxDepth</code></td> <td>-1</td> <td>If this option is specified to negative or is set to 0, recursive fields are not permitted. Setting it to 1 drops all recursive fields, 2 allows recursive fields to be recursed once, and 3 allows it to be recursed twice and so on, up to 15. Values larger than 15 are not allowed in order to avoid inadvertently creating very large schemas. If an avro message has depth beyond this limit, the Spark struct returned is truncated after the recursion limit. An example of usage can be found in section <a href="#handling-circular-references-of-avro-fields">Handling circular references of Avro fields</a></td> <td>read</td> <td>4.0.0</td> </tr> </table>

Configuration

Configuration of Avro can be done via spark.conf.set or by running SET key=value commands using SQL.

<table class="spark-config"> <thead><tr><th>Property Name</th><th>Default</th><th>Meaning</th><th>Since Version</th></tr></thead> <tr> <td>spark.sql.legacy.replaceDatabricksSparkAvro.enabled</td> <td>true</td> <td> If it is set to true, the data source provider <code>com.databricks.spark.avro</code> is mapped to the built-in but external Avro data source module for backward compatibility.

<b>Note:</b> the SQL config has been deprecated in Spark 3.2 and might be removed in the future. </td> <td>2.4.0</td>

</tr> <tr> <td>spark.sql.avro.compression.codec</td> <td>snappy</td> <td> Compression codec used in writing of AVRO files. Supported codecs: uncompressed, deflate, snappy, bzip2, xz and zstandard. Default codec is snappy. </td> <td>2.4.0</td> </tr> <tr> <td>spark.sql.avro.deflate.level</td> <td>-1</td> <td> Compression level for the deflate codec used in writing of AVRO files. Valid value must be in the range of from 1 to 9 inclusive or -1. The default value is -1 which corresponds to 6 level in the current implementation. </td> <td>2.4.0</td> </tr> <tr> <td>spark.sql.avro.xz.level</td> <td>6</td> <td> Compression level for the xz codec used in writing of AVRO files. Valid value must be in the range of from 1 to 9 inclusive. The default value is 6 in the current implementation. </td> <td>4.0.0</td> </tr> <tr> <td>spark.sql.avro.zstandard.level</td> <td>3</td> <td> Compression level for the zstandard codec used in writing of AVRO files. The default value is 3 in the current implementation. </td> <td>4.0.0</td> </tr> <tr> <td>spark.sql.avro.zstandard.bufferPool.enabled</td> <td>false</td> <td> If true, enable buffer pool of ZSTD JNI library when writing of AVRO files. </td> <td>4.0.0</td> </tr> <tr> <td>spark.sql.avro.datetimeRebaseModeInRead</td> <td><code>EXCEPTION</code></td> <td>The rebasing mode for the values of the <code>date</code>, <code>timestamp-micros</code>, <code>timestamp-millis</code> logical types from the Julian to Proleptic Gregorian calendar: </tr> <tr> <td>spark.sql.avro.datetimeRebaseModeInWrite</td> <td><code>EXCEPTION</code></td> <td>The rebasing mode for the values of the <code>date</code>, <code>timestamp-micros</code>, <code>timestamp-millis</code> logical types from the Proleptic Gregorian to Julian calendar: </tr> <tr> <td>spark.sql.avro.filterPushdown.enabled</td> <td>true</td> <td> When true, enable filter pushdown to Avro datasource. </td> <td>3.1.0</td> </tr> </table>

Compatibility with Databricks spark-avro

This Avro data source module is originally from and compatible with Databricks's open source repository spark-avro.

By default with the SQL configuration spark.sql.legacy.replaceDatabricksSparkAvro.enabled enabled, the data source provider com.databricks.spark.avro is mapped to this built-in Avro module. For the Spark tables created with Provider property as com.databricks.spark.avro in catalog meta store, the mapping is essential to load these tables if you are using this built-in Avro module.

Note in Databricks's spark-avro, implicit classes AvroDataFrameWriter and AvroDataFrameReader were created for shortcut function .avro(). In this built-in but external module, both implicit classes are removed. Please use .format("avro") in DataFrameWriter or DataFrameReader instead, which should be clean and good enough.

If you prefer using your own build of spark-avro jar file, you can simply disable the configuration spark.sql.legacy.replaceDatabricksSparkAvro.enabled, and use the option --jars on deploying your applications. Read the Advanced Dependency Management section in the Application Submission Guide for more details.

Supported types for Avro -> Spark SQL conversion

Currently Spark supports reading all primitive types and complex types under records of Avro.

<table> <thead><tr><th><b>Avro type</b></th><th><b>Spark SQL type</b></th></tr></thead> <tr> <td>boolean</td> <td>BooleanType</td> </tr> <tr> <td>int</td> <td>IntegerType</td> </tr> <tr> <td>long</td> <td>LongType</td> </tr> <tr> <td>float</td> <td>FloatType</td> </tr> <tr> <td>double</td> <td>DoubleType</td> </tr> <tr> <td>string</td> <td>StringType</td> </tr> <tr> <td>enum</td> <td>StringType</td> </tr> <tr> <td>fixed</td> <td>BinaryType</td> </tr> <tr> <td>bytes</td> <td>BinaryType</td> </tr> <tr> <td>record</td> <td>StructType</td> </tr> <tr> <td>array</td> <td>ArrayType</td> </tr> <tr> <td>map</td> <td>MapType</td> </tr> <tr> <td>union</td> <td>See below</td> </tr> </table>

In addition to the types listed above, it supports reading union types. The following three types are considered basic union types:

1. union(int, long) will be mapped to LongType.
2. union(float, double) will be mapped to DoubleType.
3. union(something, null), where something is any supported Avro type. This will be mapped to the same Spark SQL type as that of something, with nullable set to true. All other union types are considered complex. They will be mapped to StructType where field names are member0, member1, etc., in accordance with members of the union. This is consistent with the behavior when converting between Avro and Parquet.

It also supports reading the following Avro logical types:

<table> <thead><tr><th><b>Avro logical type</b></th><th><b>Avro type</b></th><th><b>Spark SQL type</b></th></tr></thead> <tr> <td>date</td> <td>int</td> <td>DateType</td> </tr> <tr> <td>timestamp-millis</td> <td>long</td> <td>TimestampType</td> </tr> <tr> <td>timestamp-micros</td> <td>long</td> <td>TimestampType</td> </tr> <tr> <td>decimal</td> <td>fixed</td> <td>DecimalType</td> </tr> <tr> <td>decimal</td> <td>bytes</td> <td>DecimalType</td> </tr> </table> At the moment, it ignores docs, aliases and other properties present in the Avro file.

Supported types for Spark SQL -> Avro conversion

Spark supports writing of all Spark SQL types into Avro. For most types, the mapping from Spark types to Avro types is straightforward (e.g. IntegerType gets converted to int); however, there are a few special cases which are listed below:

<table> <thead><tr><th><b>Spark SQL type</b></th><th><b>Avro type</b></th><th><b>Avro logical type</b></th></tr></thead> <tr> <td>ByteType</td> <td>int</td> <td></td> </tr> <tr> <td>ShortType</td> <td>int</td> <td></td> </tr> <tr> <td>BinaryType</td> <td>bytes</td> <td></td> </tr> <tr> <td>DateType</td> <td>int</td> <td>date</td> </tr> <tr> <td>TimestampType</td> <td>long</td> <td>timestamp-micros</td> </tr> <tr> <td>DecimalType</td> <td>fixed</td> <td>decimal</td> </tr> </table>

You can also specify the whole output Avro schema with the option avroSchema, so that Spark SQL types can be converted into other Avro types. The following conversions are not applied by default and require user specified Avro schema:

<table> <thead><tr><th><b>Spark SQL type</b></th><th><b>Avro type</b></th><th><b>Avro logical type</b></th></tr></thead> <tr> <td>BinaryType</td> <td>fixed</td> <td></td> </tr> <tr> <td>StringType</td> <td>enum</td> <td></td> </tr> <tr> <td>TimestampType</td> <td>long</td> <td>timestamp-millis</td> </tr> <tr> <td>DecimalType</td> <td>bytes</td> <td>decimal</td> </tr> </table>

Handling circular references of Avro fields

In Avro, a circular reference occurs when the type of a field is defined in one of the parent records. This can cause issues when parsing the data, as it can result in infinite loops or other unexpected behavior. To read Avro data with schema that has circular reference, users can use the recursiveFieldMaxDepth option to specify the maximum number of levels of recursion to allow when parsing the schema. By default, Spark Avro data source will not permit recursive fields by setting recursiveFieldMaxDepth to -1. However, you can set this option to 1 to 15 if needed.

Setting recursiveFieldMaxDepth to 1 drops all recursive fields, setting it to 2 allows it to be recursed once, and setting it to 3 allows it to be recursed twice. A recursiveFieldMaxDepth value greater than 15 is not allowed, as it can lead to performance issues and even stack overflows.

SQL Schema for the below Avro message will vary based on the value of recursiveFieldMaxDepth.

<div data-lang="avro" markdown="1"> <div class="d-none"> This div is only used to make markdown editor/viewer happy and does not display on web </div> </div>