# 1. 添加YUM仓库
yum install -y yum-utils
yum-config-manager –add-repo https://packages.clickhouse.com/rpm/stable/clickhouse.repo

# 2. 安装ClickHouse
yum install -y clickhouse-server clickhouse-client

# 3. 配置
cp /etc/clickhouse-server/config.xml /etc/clickhouse-server/config.xml.bak
cp /etc/clickhouse-server/users.xml /etc/clickhouse-server/users.xml.bak

# 4. 修改配置文件
vi /etc/clickhouse-server/config.xml
# 修改path为/bigdata/fgdata/clickhouse
# 配置remote_servers

# 5. 创建目录
mkdir -p /bigdata/fgdata/clickhouse
mkdir -p /bigdata/fgdata/clickhouse/tmp
mkdir -p /bigdata/fgdata/clickhouse/user_files
mkdir -p /bigdata/fgdata/logs/clickhouse
chown -R clickhouse:clickhouse /bigdata/fgdata/clickhouse
chown -R clickhouse:clickhouse /bigdata/fgdata/logs/clickhouse

# 6. 启动服务
systemctl start clickhouse-server
systemctl enable clickhouse-server

# 7. 验证安装
clickhouse-client
🙂 SELECT version();

# 8. 集群验证
🙂 SELECT * FROM system.clusters;

# 连接ClickHouse
clickhouse-client

# 创建数据库
🙂 CREATE DATABASE IF NOT EXISTS fgedu_db;
🙂 USE fgedu_db;

# 创建本地表（MergeTree）
🙂 CREATE TABLE IF NOT EXISTS fgedu_user_log_local (
dt Date,
ts DateTime,
user_id UInt64,
event_type String,
page String,
duration UInt32
)
ENGINE = MergeTree()
PARTITION BY toYYYYMM(dt)
ORDER BY (dt, ts, user_id)
SETTINGS index_granularity = 8192;

# 创建分布式表
🙂 CREATE TABLE IF NOT EXISTS fgedu_user_log (
dt Date,
ts DateTime,
user_id UInt64,
event_type String,
page String,
duration UInt32
)
ENGINE = Distributed(‘fgedu_cluster’, ‘fgedu_db’, ‘fgedu_user_log_local’, rand());

# 导入CSV文件
🙂 INSERT INTO fgedu_user_log
FROM INFILE ‘/bigdata/fgdata/user_log.csv’
FORMAT CSV;

# 使用clickhouse-client导入
cat /bigdata/fgdata/user_log.csv | clickhouse-client -h fgedu-ck \
–query=”INSERT INTO fgedu_db.fgedu_user_log FORMAT CSV”

# 并行导入
for file in /bigdata/fgdata/data_part_*.csv; do
cat $file | clickhouse-client -h fgedu-ck \
–query=”INSERT INTO fgedu_db.fgedu_user_log FORMAT CSV” &
done
wait

# Spark ClickHouse Connector
<dependency>
<groupId>com.github.housepower</groupId>
<artifactId>clickhouse-native-jdbc</artifactId>
<version>2.6.5</version>
</dependency>

# Spark写入ClickHouse
import org.apache.spark.sql.SparkSession
import java.util.Properties

object FgeduSpark2ClickHouse {
def main(args: Array[String]): Unit = {
val spark = SparkSession.builder()
.appName(“FgeduSpark2ClickHouse”)
.enableHiveSupport()
.getOrCreate()

val df = spark.read.table(“hive_db.fgedu_user_table”)

val properties = new Properties()
properties.setProperty(“user”, “default”)
properties.setProperty(“password”, “”)
properties.setProperty(“driver”, “com.github.housepower.jdbc.ClickHouseDriver”)
properties.setProperty(“socket_timeout”, “300000”)
properties.setProperty(“rewriteBatchedStatements”, “true”)
properties.setProperty(“batchsize”, “100000”)

df.write
.mode(“append”)
.option(“isolationLevel”, “NONE”)
.option(“numPartitions”, “10”)
.jdbc(“jdbc:clickhouse://fgedu-ck:8123/fgedu_db”,
“fgedu_user_table”, properties)

spark.stop()
}
}

# 使用ClickHouse表函数直接读取HDFS
🙂 INSERT INTO fgedu_user_log
SELECT * FROM hdfs(
‘hdfs://fgedu-nn:8020/bigdata/fgdata/hive/warehouse/fgedu_user/*’,
‘CSV’,
‘dt Date, ts DateTime, user_id UInt64, event_type String, page String, duration UInt32’
);