Details
-
Improvement
-
Status: In Progress
-
Major
-
Resolution: Unresolved
-
None
-
None
Description
Motivation
When we synchronize data from mysql to HBase, we find that when deleting data from mysql, HBase cannot delete all versions, which leads to incorrect semantics. So we want to add a parameter to control deleting the latest version or deleting all versions.
Usage
The test code is as follows.
package com.bruce; import org.apache.flink.configuration.Configuration; import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment; import org.apache.flink.table.api.EnvironmentSettings; import org.apache.flink.table.api.TableConfig; import org.apache.flink.table.api.bridge.java.StreamTableEnvironment; import static org.apache.flink.configuration.ConfigConstants.LOCAL_START_WEBSERVER; public class KafkaToHBase { public static void main(String[] args) { Configuration cfg = new Configuration(); cfg.setBoolean(LOCAL_START_WEBSERVER, true); StreamExecutionEnvironment env = StreamExecutionEnvironment.createLocalEnvironmentWithWebUI(cfg); env.setParallelism(1); EnvironmentSettings envSettings = EnvironmentSettings.newInstance() .inStreamingMode() .build(); StreamTableEnvironment tEnv = StreamTableEnvironment.create(env, envSettings); // TableConfig config = tEnv.getConfig(); // config.setIdleStateRetention(Duration.ofHours(2)); String source = "CREATE TEMPORARY TABLE IF NOT EXISTS kafka_llspay_bundles(\n" + " id STRING,\n" + " category_id STRING,\n" + " upc STRING,\n" + " `name` STRING,\n" + " price_cents STRING,\n" + " original_price_cents STRING,\n" + " short_desc STRING,\n" + " desc STRING,\n" + " cover_url STRING,\n" + " created_at STRING,\n" + " updated_at STRING,\n" + " deleted_at STRING,\n" + " extra STRING,\n" + " status STRING,\n" + " scholarship_cents STRING,\n" + " is_payback STRING,\n" + " is_support_iap STRING,\n" + " iap_product_id STRING,\n" + " neo_product_code STRING,\n" + " paid_redirect_url STRING,\n" + " subscription_type STRING\n" + ") WITH (\n" + " 'connector' = 'kafka',\n" + " 'topic' = 'dim-bundles',\n" + " 'properties.bootstrap.servers' = 'localhost:9092',\n" + " 'properties.group.id' = 'vvp_dev',\n" + " 'scan.startup.mode' = 'latest-offset',\n" + " 'value.debezium-json.schema-include' = 'true',\n" + " 'value.format' = 'debezium-json',\n" + " 'value.debezium-json.ignore-parse-errors' = 'true'\n" + ")"; String sink = "CREATE TEMPORARY TABLE IF NOT EXISTS dim_hbase (\n" + " rowkey STRING,\n" + " cf ROW<id STRING, category_id STRING, upc STRING, `name` STRING, price_cents STRING, original_price_cents STRING, short_desc STRING, `desc` STRING, cover_url STRING, created_at STRING, updated_at STRING, deleted_at STRING, extra STRING, status STRING, scholarship_cents STRING, is_payback STRING, is_support_iap STRING, iap_product_id STRING, neo_product_code STRING, paid_redirect_url STRING, subscription_type STRING>\n" + ") with (\n" + " 'connector'='hbase-2.2',\n" + " 'table-name'='dim_hbase',\n" + " 'zookeeper.quorum'='localhost:2181',\n" + " 'sink.buffer-flush.max-size' = '0',\n" + " 'sink.buffer-flush.max-rows' = '1',\n" + " 'sink.delete.mode' = 'all-versions'\n" + ")"; String dml = "INSERT INTO dim_hbase\n" + "SELECT \n" + " upc as rowkey,\n" + " ROW(\n" + " id, category_id, upc, `name`, price_cents, original_price_cents, short_desc, `desc` , cover_url , created_at, updated_at, deleted_at, extra , status , scholarship_cents , is_payback , is_support_iap , iap_product_id , neo_product_code , paid_redirect_url , subscription_type)\n" + "FROM kafka_llspay_bundles"; tEnv.executeSql(source); tEnv.executeSql(sink); tEnv.executeSql(dml); } }
Test
- After the test, we found that the deleted CF was the CF specified in the Flink SQL DML statement, without affecting other CF.
Two CFs to one CF.
The data of cf2 is still there. So the deleted CF was the CF specified in the Flink SQL DML statement, without affecting other CF.
Reference
Please look at the following task link.
Attachments
Attachments
Issue Links
- duplicates
-
FLINK-25330 Flink SQL doesn't retract all versions of Hbase data
-
- Closed
-
- fixes
-
FLINK-25330 Flink SQL doesn't retract all versions of Hbase data
-
- Closed
-
- links to