druid/docs/content/development/extensions-contrib/parquet.md

91 lines
3.0 KiB
Markdown

# Parquet
This extension enables Druid to ingest and understand the Apache Parquet data format offline.
## Parquet Hadoop Parser
This is for batch ingestion using the HadoopDruidIndexer. The inputFormat of inputSpec in ioConfig must be set to `"io.druid.data.input.parquet.DruidParquetInputFormat"`. Make sure also to include "io.druid.extensions:druid-avro-extensions" as an extension.
Field | Type | Description | Required
----------|-------------|----------------------------------------------------------------------------------------|---------
type | String | This should say `parquet` | yes
parseSpec | JSON Object | Specifies the timestamp and dimensions of the data. Should be a timeAndDims parseSpec. | yes
For example:
```json
{
"type": "index_hadoop",
"spec": {
"ioConfig": {
"type": "hadoop",
"inputSpec": {
"type": "static",
"inputFormat": "io.druid.data.input.parquet.DruidParquetInputFormat",
"paths": "no_metrics"
},
"metadataUpdateSpec": {
"type": "postgresql",
"connectURI": "jdbc:postgresql://localhost/druid",
"user" : "druid",
"password" : "asdf",
"segmentTable": "druid_segments"
},
"segmentOutputPath": "tmp/segments"
},
"dataSchema": {
"dataSource": "no_metrics",
"parser": {
"type": "parquet",
"parseSpec": {
"format": "timeAndDims",
"timestampSpec": {
"column": "time",
"format": "auto"
},
"dimensionsSpec": {
"dimensions": [
"name"
],
"dimensionExclusions": [],
"spatialDimensions": []
}
}
},
"metricsSpec": [{
"type": "count",
"name": "count"
}],
"granularitySpec": {
"type": "uniform",
"segmentGranularity": "DAY",
"queryGranularity": "ALL",
"intervals": ["2015-12-31/2016-01-02"]
}
},
"tuningConfig": {
"type": "hadoop",
"workingPath": "tmp/working_path",
"partitionsSpec": {
"targetPartitionSize": 5000000
},
"jobProperties" : {},
"leaveIntermediate": true
}
}
}
```
Almost all the fields listed above are required, including `inputFormat`, `metadataUpdateSpec`(`type`, `connectURI`, `user`, `password`, `segmentTable`). Set `jobProperties` to make hdfs path timezone unrelated.
It is no need to make your cluster to update to SNAPSHOT, you can just fire a hadoop job with your local compiled jars like:
```bash
HADOOP_CLASS_PATH=`hadoop classpath | sed s/*.jar/*/g`
java -Xmx32m -Duser.timezone=UTC -Dfile.encoding=UTF-8 \
-classpath config/overlord:config/_common:lib/*:$HADOOP_CLASS_PATH:extensions/druid-avro-extensions/* \
io.druid.cli.Main index hadoop \
wikipedia_hadoop_parquet_job.json
```