More information about compilation and usage, please visit Spark Doris Connector
You need to copy customer_env.sh.tpl to customer_env.sh before build and you need to configure it before build.
git clone git@github.com:apache/doris-spark-connector.git
cd doris-spark-connector/spark-doris-connector
./build.sh- download and compile Spark Doris Connector from https://github.com/apache/doris-spark-connector, we suggest compile Spark Doris Connector by Doris official image。
$ docker pull apache/doris:build-env-ldb-toolchain-latest-
the result of compile jar is like：spark-doris-connector-3.1_2.12-1.0.0-SNAPSHOT.jar
-
download spark for https://spark.apache.org/downloads.html .if in china there have a good choice of tencent link https://mirrors.cloud.tencent.com/apache/spark/spark-3.1.2/
#download
wget https://mirrors.cloud.tencent.com/apache/spark/spark-3.1.2/spark-3.1.2-bin-hadoop3.2.tgz
#decompression
tar -xzvf spark-3.1.2-bin-hadoop3.2.tgz- config Spark environment
vim /etc/profile
export SPARK_HOME=/your_parh/spark-3.1.2-bin-hadoop3.2
export PATH=$PATH:$SPARK_HOME/bin
source /etc/profile- copy spark-doris-connector-3.1_2.12-1.0.0-SNAPSHOT.jar to spark jars directory。
cp /your_path/spark-doris-connector/target/spark-doris-connector-3.1_2.12-1.0.0-SNAPSHOT.jar $SPARK_HOME/jars-
created doris database and table。
create database mongo_doris; use mongo_doris; CREATE TABLE data_sync_test_simple ( _id VARCHAR(32) DEFAULT '', id VARCHAR(32) DEFAULT '', user_name VARCHAR(32) DEFAULT '', member_list VARCHAR(32) DEFAULT '' ) DUPLICATE KEY(_id) DISTRIBUTED BY HASH(_id) BUCKETS 10 PROPERTIES("replication_num" = "1"); INSERT INTO data_sync_test_simple VALUES ('1','1','alex','123');
- Input this coed in spark-shell.
import org.apache.doris.spark._
val dorisSparkRDD = sc.dorisRDD(
tableIdentifier = Some("mongo_doris.data_sync_test"),
cfg = Some(Map(
"doris.fenodes" -> "127.0.0.1:8030",
"doris.request.auth.user" -> "root",
"doris.request.auth.password" -> ""
))
)
dorisSparkRDD.collect()- mongo_doris:doris database name
- data_sync_test:doris table mame.
- doris.fenodes:doris FE IP:http_port
- doris.request.auth.user:doris user name.
- doris.request.auth.password:doris password
- if Spark is Cluster model,upload Jar to HDFS，add doris-spark-connector jar HDFS URL in spark.yarn.jars.
spark.yarn.jars=hdfs:///spark-jars/doris-spark-connector-3.1.2-2.12-1.0.0.jarLink：apache/doris#9486
- in pyspark,input this code in pyspark shell command.
dorisSparkDF = spark.read.format("doris")
.option("doris.table.identifier", "mongo_doris.data_sync_test")
.option("doris.fenodes", "127.0.0.1:8030")
.option("user", "root")
.option("password", "")
.load()
# show 5 lines data
dorisSparkDF.show(5)Use a three-part catalog.database.table identifier to read tables from Doris
external catalogs such as Hive, Iceberg, and Paimon:
spark.read.format("doris")
.option("doris.fenodes", "127.0.0.1:8030")
.option("doris.table.identifier", "hive_catalog.database.table")
.option("doris.user", "root")
.option("doris.password", "")
.load()Arrow Flight SQL must be enabled on Doris. The connector discovers its port;
when frontend discovery is disabled, set doris.read.arrow-flight-sql.port
explicitly. Three-part identifiers require Arrow mode and do not fall back to
Thrift. The same identifier works with SQL USING doris and the RDD API.
External reads use one Spark partition. Doris plans and executes the external
scan. Column projection and filter pushdown are retained. Complex types
use the existing connector conversions: ARRAY as a string (or an array of
strings with doris.read.array.native-type=true), MAP as a map of strings, and
STRUCT as a string. Writing through a three-part identifier is not supported.
The connector supports one-way TLS for Doris HTTP APIs and Stream Load, MySQL/JDBC metadata queries, BE Thrift reads, and Arrow Flight SQL reads. TLS is disabled by default.
spark.read.format("doris")
.option("doris.fenodes", "fe.example.com:8040")
.option("doris.table.identifier", "database.table")
.option("user", "root")
.option("password", "")
.option("doris.enable.tls", "true")
.option("doris.tls.ca-certificate-path", "/etc/doris-tls/ca.pem")
.load()| Option | Default | Description |
|---|---|---|
doris.enable.tls |
false |
Enables TLS for all Doris protocols not listed in doris.tls.excluded-protocols. |
doris.tls.ca-certificate-path |
empty | Path to a PEM file containing one or more trusted X.509 CA certificates. An empty value uses the JVM or client-library default trust roots. |
doris.tls.skip-hostname-verification |
false |
Disables hostname verification while retaining CA verification. This mode is not supported by Arrow Flight SQL. |
doris.tls.excluded-protocols |
empty | Comma-separated plaintext exceptions: http, mysql, thrift, and arrowflight. |
The connector validates certificate chains and hostnames by default and does not modify JVM-global TLS settings. Client certificates and mutual TLS are not supported.
The CA file must be available at the same path on the Spark driver and every
executor. For YARN, distribute it with --files /local/path/ca.pem#ca.pem and
configure doris.tls.ca-certificate-path=ca.pem. For Kubernetes, mount the CA
from a Secret or volume into both driver and executor pods and configure the
mounted path.
| doris | spark |
|---|---|
| BOOLEAN | BooleanType |
| TINYINT | ByteType |
| SMALLINT | ShortType |
| INT | IntegerType |
| BIGINT | LongType |
| LARGEINT | StringType |
| FLOAT | FloatType |
| DOUBLE | DoubleType |
| DECIMAL(M,D) | DecimalType(M,D) |
| DATE | DateType |
| DATETIME | TimestampType |
| CHAR(L) | StringType |
| VARCHAR(L) | StringType |
| STRING | StringType |
| ARRAY | ARRAY |
| MAP | MAP |
| STRUCT | STRUCT |
If you find any bugs, feel free to file a GitHub issue or fix it by submitting a pull request.
Contact us through the following mailing list.
| Name | Scope | |||
|---|---|---|---|---|
| dev@doris.apache.org | Development-related discussions | Subscribe | Unsubscribe | Archives |
- Doris official site - https://doris.apache.org
- Developer Mailing list - dev@doris.apache.org. Mail to dev-subscribe@doris.apache.org, follow the reply to subscribe the mail list.
- Slack channel - Join the Slack