{"id":10008274,"date":"2023-09-18T20:28:09","date_gmt":"2023-09-18T12:28:09","guid":{"rendered":"https:\/\/blog.itmax.cn\/?p=10008274"},"modified":"2023-09-18T20:28:09","modified_gmt":"2023-09-18T12:28:09","slug":"smartnews-%e5%9f%ba%e4%ba%8e-flink-%e7%9a%84-iceberg-%e5%ae%9e%e6%97%b6%e6%95%b0%e6%8d%ae%e6%b9%96%e5%ae%9e%e8%b7%b5","status":"publish","type":"post","link":"https:\/\/itmax.cn\/?p=10008274","title":{"rendered":"SmartNews \u57fa\u4e8e Flink \u7684 Iceberg \u5b9e\u65f6\u6570\u636e\u6e56\u5b9e\u8df5"},"content":{"rendered":"<blockquote>\n<p>\u6458\u8981\uff1a\u672c\u6587\u6574\u7406\u81ea SmartNews \u6570\u636e\u5e73\u53f0\u67b6\u6784\u5e08 Apache Iceberg Contributor \u6222\u6e05\u96e8\uff0c\u5728 Flink Forward Asia 2022 \u5b9e\u65f6\u6e56\u4ed3\u4e13\u573a\u7684\u5206\u4eab\u3002\u672c\u7bc7\u5185\u5bb9\u4e3b\u8981\u5206\u4e3a\u4e94\u4e2a\u90e8\u5206\uff1a<\/p>\n<ol>\n<li>SmartNews \u6570\u636e\u6e56\u4ecb\u7ecd<\/li>\n<li>\u57fa\u4e8e Icebergv1 \u683c\u5f0f\u7684\u6570\u636e\u6e56\u5b9e\u8df5<\/li>\n<li>\u57fa\u4e8e Flink \u5b9e\u65f6\u66f4\u65b0\u7684\u6570\u636e\u6e56\uff08Iceberg v2\uff09\u89e3\u51b3\u65b9\u6848<\/li>\n<li>\u5b9e\u65f6\u66f4\u65b0\u5c0f\u6587\u4ef6\u95ee\u9898\u7684\u4f18\u5316<\/li>\n<li>\u603b\u7ed3\u4e0e\u5c55\u671b<\/li>\n<\/ol>\n<\/blockquote>\n<p><a target=\"_blank\" href=\"https:\/\/link.segmentfault.com\/?enc=P%2BjQYWGeGBLbevQCpIlM0w%3D%3D.4o4QfuGitaS%2BU%2FE7M16AJy2XBnMnrAF1t%2BigfBpGeutqAcIyXoqBj541W5I31d6SRpSwSpu5tub%2BhWHFU0zKSvw6857PM8hEDu1eJTXm5SDi3KPLsz1YL5yBILoGmuGSDGM%2BnliTRfQJ4tJu6pk2ig%3D%3D\" rel=\"noopener noreferrer\">\u70b9\u51fb\u67e5\u770b\u539f\u6587\u89c6\u9891 &#038; \u6f14\u8bb2PPT<\/a><\/p>\n<h2>\u4e00\u3001SmartNews \u6570\u636e\u6e56\u4ecb\u7ecd<\/h2>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228632\" alt=\"1\" title=\"1\"><\/p>\n<p>2012 \u5e74\uff0cSmartNews \u516c\u53f8\u5728\u65e5\u672c\u4e1c\u4eac\u6210\u7acb\u3002\u4e00\u76f4\u4e13\u6ce8\u4e8e PGC \u65b0\u95fb\uff0c\u662f\u4e00\u6b3e\u5728\u65e5\u672c\u5904\u4e8e\u9886\u5148\u5730\u4f4d\u7684\u65b0\u95fb APP\u3002\u76ee\u524d\uff0c\u670d\u52a1\u7684\u5ba2\u6237\u4e3b\u8981\u96c6\u4e2d\u5728\u65e5\u672c\u3001\u6b27\u7f8e\u7b49\u56fd\u5bb6\u3002SmartNews \u516c\u53f8\u5728\u65e5\u672c\u3001\u7f8e\u56fd\u548c\u4e2d\u56fd\u5747\u8bbe\u6709\u529e\u516c\u5ba4\uff0c\u5728 2019 \u5e74\u5165\u9a7b\u5317\u4eac\u548c\u4e0a\u6d77\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228633\" alt=\"2\" title=\"2\" loading=\"lazy\"><\/p>\n<p>SmartNews \u6570\u636e\u6e56\u4e3b\u8981\u5b58\u50a8\u6240\u6709\u5e7f\u544a\u6570\u636e\uff0c\u5305\u62ec\u4ece\u670d\u52a1\u5668\u7aef\u6536\u96c6\u5230\u7684\u70b9\u51fb\/\u8f6c\u5316\u7b49\u4e8b\u4ef6\u4fe1\u606f\uff0c\u7ef4\u8868\u4fe1\u606f\u3002\u5176\u4e3b\u8981\u7684\u5e7f\u544a\u4fe1\u606f\u90fd\u5b58\u50a8\u5728 Kafka \u4e0a\uff0c\u670d\u52a1\u5668\u7aef\u5728\u6536\u96c6\u5230\u4e8b\u4ef6\u540e\uff0c\u4f1a\u76f4\u63a5\u5b9e\u65f6\u5199\u5165 Kafka\u3002\u5176\u4ed6\u7684\u7ef4\u8868\u4fe1\u606f\uff0c\u6bd4\u5982\u5e7f\u544a\u4fe1\u606f\u3001\u7edf\u8ba1\u4fe1\u606f\u7b49\u7b49\uff0c\u4e3b\u8981\u5b58\u50a8\u5728 MySQL \u6216 Hive \u4e2d\u3002\u8fd9\u4e9b\u4fe1\u606f\u4e00\u822c\u4ee5\u5b9e\u65f6\u6216\u5c0f\u65f6\u7ea7\u522b\u66f4\u65b0\u3002<\/p>\n<p>\u6570\u636e\u6e56\u7684\u4e0b\u6e38\u662f\u4e1a\u52a1\u7aef\u7684 ETL \u6216\u5b9e\u65f6\u62a5\u8868\u6570\u636e\uff0c\u662f\u4e0b\u6e38\u6570\u636e\u7684\u7edf\u4e00\u5165\u53e3\u3002\u56e0\u6b64\uff0c\u6211\u4eec\u5c3d\u91cf\u628a\u6240\u6709\u7ef4\u5ea6\u90fd\u653e\u8fdb\u6765\uff0c\u505a\u6210\u4e00\u4e2a\u5927\u5bbd\u8868\uff0c\u4f9b\u4e0b\u6e38\u5b9e\u65f6\u67e5\u8be2\u4f7f\u7528\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228634\" alt=\"3\" title=\"3\" loading=\"lazy\"><\/p>\n<p>\u63a5\u4e0b\u6765\u4ecb\u7ecd\u4e0b\u6570\u636e\u6e56\u9700\u8981\u89e3\u51b3\u7684\u6280\u672f\u6311\u6218\u3002<\/p>\n<ul>\n<li>\u7b2c\u4e00\uff0c\u6309\u7167\u5e7f\u544a\u4e3b\u952e\u53bb\u91cd\u3002\u4e0a\u6e38\u6570\u636e\u6309\u7167\u6bcf\u6761\u5e7f\u544a\u7684\u4e8b\u4ef6\uff0c\u8fdb\u884c\u6536\u96c6\u3002\u6bd4\u5982\u4e00\u6761\u5e7f\u544a\u7684\u70b9\u51fb\u6216\u8005\u8f6c\u5316\u4f1a\u751f\u6210\u591a\u6761\u8bb0\u5f55\uff0c\u56e0\u6b64\u6211\u4eec\u9700\u8981\u5c06\u8fd9\u4e9b\u4e8b\u4ef6\u6253\u5e73\u3002\u5176\u6b21\u662f\u4e0a\u6e38\u7684 Kafka \u6570\u636e\uff0c\u53ef\u80fd\u5305\u542b\u4e86\u4e00\u5b9a\u7a0b\u5ea6\u7684\u91cd\u590d\u6570\u636e\u3002<\/li>\n<li>\u7b2c\u4e8c\uff0c\u9700\u8981\u66f4\u65b0\u70b9\u51fb\/\u8f6c\u5316\u65f6\u95f4\u6233\u5b57\u6bb5\u3002\u6bd4\u5982\u4e8b\u4ef6\u7684\u65f6\u95f4\u6233\uff0c\u9700\u8981\u8ba1\u7b97\u6700\u65b0\u4e00\u6b21\u7684\u65f6\u95f4\uff0c\u9700\u8981\u5bf9\u6570\u636e\u6e56\u6267\u884c\u66f4\u65b0\u64cd\u4f5c\u3002<\/li>\n<li>\u7b2c\u4e09\uff0c\u4e0b\u6e38\u8fd1\u5b9e\u65f6\u8bfb\u53d6\u3002\u8981\u6c42\u6570\u636e\u6e56\u5177\u6709\u540c\u65f6\u5199\u5165\/\u8bfb\u53d6\u7684\u64cd\u4f5c\u3002\u800c Hive \u5728\u91cd\u5199\u6570\u636e\u7684\u8fc7\u7a0b\u4e2d\u662f\u4f1a\u5f71\u54cd\u5230\u4e0b\u6e38\u6b63\u5728\u53d1\u751f\u7684\u67e5\u8be2\uff0c\u8fd9\u5c31\u8981\u6c42\u6211\u4eec\u9700\u8981\u4e00\u4e2a\u65b0\u7684\u89e3\u51b3\u65b9\u6848\u3002<\/li>\n<\/ul>\n<h2>\u4e8c\u3001\u57fa\u4e8e Icebergv1 \u683c\u5f0f\u7684\u6570\u636e\u6e56\u5b9e\u8df5<\/h2>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228635\" alt=\"4\" title=\"4\" loading=\"lazy\"><\/p>\n<p>\u4e0a\u56fe\u662f\u6211\u4eec\u7b2c\u4e00\u4e2a\u89e3\u51b3\u65b9\u6848\u7684\u6574\u4f53\u67b6\u6784\u3002\u5728\u8fd9\u4e2a\u89e3\u51b3\u65b9\u6848\u4e2d\uff0c\u6211\u4eec\u91c7\u7528\u4e86 Spark \u8ba1\u7b97\u5f15\u64ce\uff0c\u628a\u6240\u6709\u7684\u5e7f\u544a\u4e8b\u4ef6\u6309\u7167\u4e3b\u952e\u8fdb\u884c\u6253\u5e73\u5e76\u53bb\u91cd\u3002\u7136\u540e\uff0c\u6240\u6709\u7684\u7ef4\u8868\u8fdb\u884c\u67e5\u8be2 join\u3002<\/p>\n<p>\u9664\u6b64\u4e4b\u5916\uff0c\u6211\u4eec\u5c06\u6570\u636e\u6e90\u5207\u6362\u6210 S3 \u6587\u4ef6\uff0c\u6ca1\u6709\u7528\u6d41\u5f0f\u6570\u636e\u6e90\u3002\u5176\u4e3b\u8981\u539f\u56e0\u5982\u4e0b\uff1a<\/p>\n<ul>\n<li>\u7b2c\u4e00\uff0c\u8fd9\u4e2a\u65b9\u6848\u662f\u4e00\u4e2a\u5c0f\u65f6\u7ea7\u522b\u7684\u89e3\u51b3\u65b9\u6848\uff0c\u5e76\u4e0d\u9700\u8981\u5b9e\u65f6\u8bfb\u53d6\u6d41\u5f0f\u6570\u636e\u3002<\/li>\n<li>\u7b2c\u4e8c\uff0c\u6211\u4eec\u5728\u8bbe\u8ba1 Spark \u4efb\u52a1\u65f6\uff0c\u4f1a\u5b9a\u4e49\u4e00\u4e2a\u6700\u5c0f\u7684\u6267\u884c\u5355\u5143\u3002\u5c06\u76ee\u6807\u6570\u636e\u6e90\u9650\u5236\u5728\u67d0\u4e00\u5929\u7684\u67d0\u4e00\u65f6\u95f4\u3002\u901a\u8fc7 S3 \u6587\u4ef6\u7684\u5206\u533a\u4fe1\u606f\uff0c\u5c31\u53ef\u4ee5\u76f4\u63a5\u8fdb\u884c\u8bfb\u53d6\u3002<\/li>\n<li>\u7b2c\u4e09\uff0c\u4e3a\u4e86\u964d\u4f4e\u4e00\u5b9a\u7684\u5bb9\u9519\u6982\u7387\u3002\u76ee\u524d\uff0c\u6211\u4eec\u7684\u4e1a\u52a1\u9700\u8981\u56de\u6eda\u8fc7\u53bb\u56db\u5929\u7684\u6570\u636e\u3002\u6bd4\u5982\u6709\u4e00\u4e2a\u6bd4\u8f83\u5927\u7684 Spark \u4efb\u52a1\u9700\u8981\u91cd\u5199\uff0c\u5982\u679c Spark \u4efb\u52a1\u5931\u8d25\uff0c\u4f1a\u5bfc\u81f4\u6574\u4e2a\u4efb\u52a1\u5931\u8d25\u3002\u5982\u679c\u8bbe\u8ba1\u4e3a\u6700\u5c0f\u6267\u884c\u5355\u5143\uff0c\u6bcf\u4e2a Spark \u4efb\u52a1\u53ea\u5904\u7406\u67d0\u4e2a\u5c0f\u65f6\u7684\u6570\u636e\uff0c\u5bb9\u9519\u51e0\u7387\u4f1a\u5927\u5e45\u63d0\u5347\u3002<\/li>\n<\/ul>\n<p>\u4e3a\u4e86\u907f\u514d\u4e00\u4e9b\u91cd\u590d\u8ba1\u7b97\uff0c\u6211\u4eec\u4e5f\u4f1a\u53bb\u68c0\u6d4b\u5f53\u524d\u5c0f\u65f6\u662f\u5426\u6bd4\u4e0a\u6b21 Spark \u4efb\u52a1\u542f\u52a8\u7684\u65f6\u5019\u6709\u65b0\u6dfb\u52a0\u6587\u4ef6\uff0c\u901a\u8fc7 airflow \u6765\u63a7\u5236 Spark \u4efb\u52a1\u7684\u542f\u52a8\u4e0e\u91cd\u8bd5\u3002<\/p>\n<p>\u6bcf\u4e00\u4e2a\u72ec\u7acb\u7684 Spark \u4efb\u52a1\u90fd\u4f1a\u53bb\u5c1d\u8bd5 overwrite \u67d0\u4e00\u4e2a\u5c0f\u65f6\u7684\u6570\u636e\u5230 Iceberg \u4e2d\u3002\u6eda\u52a8\u5237\u65b0\u8fc7\u53bb 4 \u5929\/96 \u4e2a\u5c0f\u65f6\u7684\u6570\u636e &#8211; \u8fd9\u4e5f\u662f\u8fd9\u4e2a\u89e3\u51b3\u65b9\u6848\u7684\u4e00\u4e2a\u9650\u5236\uff0c\u5176\u5b9e\u9645\u573a\u666f\u4e2d\u8fd9\u4e2a\u5237\u65b0\u7684\u7a97\u53e3\u7406\u8bba\u503c\u662f 30 \u5929\uff0c\u4f46\u662f\u8003\u8651\u5230\u6210\u672c\u7b49\u56e0\u7d20\uff0c\u8fd9\u4e2a\u65b9\u6848\u53ea\u56de\u5237\u8fc7\u53bb 96 \u4e2a\u5c0f\u65f6\u3002\u4e0e\u6b64\u540c\u65f6\uff0c\u4e0b\u6e38\u4e5f\u4f1a\u901a\u8fc7\u4e00\u4e9b\u67e5\u8be2\u5f15\u64ce\u6765\u5bf9\u8fd9\u4e2a\u6570\u636e\u6e56\u6570\u636e\u8fdb\u884c\u5b9e\u65f6\u67e5\u8be2\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228636\" alt=\"5\" title=\"5\" loading=\"lazy\"><\/p>\n<p>\u5728\u8fd9\u4e2a\u89e3\u51b3\u65b9\u6848\u4e2d\uff0c\u89e3\u51b3\u4e86\u4e4b\u524d\u6211\u4eec\u63d0\u5230\u7684\u4e00\u4e9b\u6311\u6218\u3002\u6bd4\u5982\u5728 Spark \u4f5c\u4e1a\u4e2d\uff0c\u6309\u7167\u4e3b\u952e\u8fdb\u884c\u53bb\u91cd\uff0c\u5e76\u4e14\u66f4\u65b0\u76f8\u5e94\u7684\u65f6\u95f4\u6233\u3002\u901a\u8fc7 Iceberg \u89e3\u51b3\u65b9\u6848\uff0c\u4e0d\u4f46\u9694\u79bb\u4e0a\u4e0b\u6e38\u7684\u8bfb\u5199\uff0c\u800c\u4e14\u63d0\u4f9b\u4e86\u5c0f\u65f6\u7ea7\u522b\u7684\u66f4\u65b0\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228637\" alt=\"6\" title=\"6\" loading=\"lazy\"><\/p>\n<p>\u4f46\u8fd9\u4e2a\u65b9\u6848\u4e5f\u6709\u5f88\u591a\u4e0d\u8db3\u3002\u6bd4\u5982\u5360\u7528 Infra \u8d44\u6e90\u592a\u591a\uff0c\u8ba1\u7b97\u8d44\u6e90\u7684\u6d6a\u8d39\u3002\u901a\u8fc7\u8ba1\u7b97\u53d1\u73b0\uff0c\u9700\u8981\u66f4\u65b0\u7684\u884c\u53ea\u5360\u603b\u4f53\u7684 1%~2%\u5de6\u53f3\u3002\u9664\u6b64\u4e4b\u5916\uff0c\u8fd8\u6709\u5b58\u50a8\u8d44\u6e90\u6d6a\u8d39\u7684\u95ee\u9898\u3002Spark \u6bcf\u6b21\u4ece overwrite \u63d0\u4ea4\u5230 Iceberg \u7684\u8fc7\u7a0b\u4e2d\uff0c\u90fd\u9700\u8981\u91cd\u5199\u6574\u4e2a\u6570\u636e\u3002\u5173\u4e8e\u5e76\u884c\u63d0\u4ea4\u5230 Iceberg \u7684\u9501\u95ee\u9898\uff0c\u6bcf\u4e2a\u6700\u5c0f\u7684 Spark \u6267\u884c\u5355\u5143\uff0c\u4f1a\u540c\u65f6\u6267\u884c\u63d0\u4ea4 Iceberg \u64cd\u4f5c\u3002\u5728\u5411 Iceberg \u63d0\u4ea4\u7684\u8fc7\u7a0b\u4e2d\uff0c\u4f1a\u5148\u4ece Hive \u91cc\u62ff\u4e00\u4e2a\u9501\uff0c\u5bfc\u81f4\u5927\u5bb6\u5bf9\u9501\u5b58\u5728\u7ade\u4e89\uff0c\u9020\u6210\u4e86\u8d44\u6e90\u7684\u6d6a\u8d39\u3002<\/p>\n<h2>\u4e09\u3001\u57fa\u4e8e Flink \u5b9e\u65f6\u66f4\u65b0\u7684\u6570\u636e\u6e56(Iceberg v2\uff09\u89e3\u51b3\u65b9\u6848<\/h2>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228638\" alt=\"7\" title=\"7\" loading=\"lazy\"><\/p>\n<p>\u6211\u4eec\u7ecf\u8fc7\u5145\u5206\u7684\u8c03\u7814\u4e4b\u540e\uff0c\u51b3\u5b9a\u91c7\u7528 Flink+Iceberg V2 \u7684\u65b9\u5f0f\uff0c\u8fdb\u884c\u5b9e\u65f6\u66f4\u65b0\u3002\u8fd9\u4e2a\u89e3\u51b3\u65b9\u6848\u5229\u7528\u4e86 Iceberg V2 \u652f\u6301\u884c\u7ea7\u522b\u66f4\u65b0\uff0c\u5176\u6b21\u662f Flink \u7684\u5b9e\u65f6\u5199\u5165\u3002\u56e0\u4e3a Flink \u5728\u5199\u5165 Iceberg \u7684\u8fc7\u7a0b\u4e2d\uff0c\u4f7f\u7528\u4e86 Merge On Read\u3002\u6240\u4ee5 Flink \u53ea\u4f1a\u5199\u5165\u9700\u8981\u66f4\u65b0\u7684\u6570\u636e\u3002<\/p>\n<p>\u7531\u4e8e\u6211\u4eec\u53ea\u6709 1%\u7684\u6570\u636e\u91cf\u9700\u8981\u66f4\u65b0\uff0c\u6240\u4ee5 Merge On Read \u6a21\u5f0f\u975e\u5e38\u9002\u5408\u5f53\u524d\u7684\u4e1a\u52a1\u573a\u666f\u3002<\/p>\n<p>\u9664\u6b64\u4e4b\u5916\uff0c\u6211\u4eec\u5e0c\u671b\u901a\u8fc7 MySQL CDC \u7684\u6d41\u5f0f\u89e3\u51b3\u65b9\u6848\uff0c\u89e3\u51b3 dimension join \u7ef4\u8868\u67e5\u8be2\uff0c\u53ef\u4ee5\u66f4\u5feb\u3001\u66f4\u51c6\u786e\u7684\u5c06\u7ef4\u8868\u4fe1\u606f\u5199\u5165\u6570\u636e\u5e93\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228639\" alt=\"8\" title=\"8\" loading=\"lazy\"><\/p>\n<p>\u5728\u8fd9\u4e2a\u65b0\u7684\u89e3\u51b3\u65b9\u6848\u4e2d\uff0c\u6211\u4eec\u5c06\u4e0a\u6e38\u7684\u6570\u636e\u6e90\u90fd\u6d41\u5f0f\u5316\uff0c\u53d1\u9001\u5230 Kafka \u4e2d\u3002\u4e0e\u6b64\u540c\u65f6\uff0c\u5c06 MySQL \u7684\u7ef4\u8868\u4fe1\u606f\u901a\u8fc7 CDC \u7684\u65b9\u5f0f\uff0c\u8f93\u5165\u5230 Flink \u4efb\u52a1\u91cc\u3002Flink \u518d\u5c06\u8fd9\u4e9b\u7ef4\u8868\u4fe1\u606f\u901a\u8fc7 broadcast \u5230 State \u4e2d\uff0c\u4f9b\u4e0b\u6e38\u67e5\u8be2\u3002<\/p>\n<p>Flink \u5728\u901a\u8fc7 Iceberg Sink \u7684 Upsert Mode \u6765\u5c06\u6570\u636e\u5b9e\u65f6\u5199\u5165\u5230\u6570\u636e\u6e56\u4e2d\u3002offline \u7684\u8bdd\uff0c\u6211\u4eec\u518d\u901a\u8fc7 airflow \u6765\u5b9a\u65f6\u542f\u52a8\u4e00\u4e9b Spark \u4efb\u52a1\u6765\u505a\u6570\u636e\u6587\u4ef6\u7684\u5408\u5e76\uff0c\u4e3b\u8981\u662f\u4e3a\u4e86\u89e3\u51b3\u5c0f\u6587\u4ef6\u7684\u95ee\u9898\uff0c\u6211\u4eec\u5728\u540e\u9762\u7684\u7ae0\u8282\u4e5f\u4f1a\u6709\u8be6\u7ec6\u4ecb\u7ecd\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228640\" alt=\"9\" title=\"9\" loading=\"lazy\"><\/p>\n<p>\u5bf9\u6bd4\u4e0a\u8ff0\u4e24\u79cd\u4e0d\u540c\u7684\u89e3\u51b3\u65b9\u6848\uff0c\u53ef\u4ee5\u770b\u51fa\u4ee5\u4e0b\u533a\u522b\u3002\u9996\u5148\uff0cSpark + Iceberg v1 \u7684\u5199\u5165\u65b9\u5f0f\u662f overwrite\u3002\u6bcf\u6b21\u4f1a\u5c06\u6240\u6709\u7684\u6570\u636e\u96c6\u91cd\u65b0\u8ba1\u7b97\uff0c\u7136\u540e\u91cd\u65b0\u653e\u5230\u6570\u636e\u6e56\u4e2d\u3002Flink + Iceberg v2 \u7684\u5199\u5165\u65b9\u5f0f\u662f Upsert\uff0c\u53ea\u662f\u5c06\u66f4\u65b0\u7684\u6570\u636e\u5199\u5165\u5230\u6570\u636e\u6e56\u4e2d\u3002<\/p>\n<p>\u4ece\u8f93\u51fa\u6587\u4ef6\u6570\u91cf\u7684\u89d2\u5ea6\u6765\u8bb2\uff0cSpark + Iceberg v1 \u7684\u6587\u4ef6\u5927\u5c0f\u53ef\u63a7\uff0c\u6570\u91cf\u53ef\u63a7\u3002\u56e0\u4e3a\u6bcf\u6b21\u8f93\u5165\u7684\u90fd\u662f\u8fd9\u4e2a\u5c0f\u65f6\u7684\u5168\u91cf\u6570\u636e\uff0c\u53ef\u4ee5\u6309\u7167\u9700\u6c42\u6765\u63a7\u5236\u6587\u4ef6\u5927\u5c0f\uff0c\u63a7\u5236\u6587\u4ef6\u6570\u91cf\u3002Flink + Iceberg v2 \u4f1a\u4ea7\u751f\u5927\u91cf\u7684\u5c0f\u6587\u4ef6\uff0c\u5e26\u6765\u5de8\u5927\u7684\u6311\u6218\u3002<\/p>\n<p>\u4ece\u8ba1\u7b97\u65b9\u5f0f\u7684\u89d2\u5ea6\u6765\u8bb2\uff0cSpark \u9700\u8981\u5168\u90e8\u91cd\u65b0\u8ba1\u7b97\u3002Flink + Iceberg v2 \u4ec5\u9700\u8981\u8ba1\u7b97\u66f4\u65b0\u7684\u6570\u636e\u3002<\/p>\n<p>\u4ece\u65f6\u6548\u6027\u7684\u89d2\u5ea6\u6765\u8bb2\uff0cSpark + Iceberg v1 \u63d0\u4f9b\u7684\u662f\u5c0f\u65f6\u7ea7\u522b\u7684\u89e3\u51b3\u65b9\u6848\u3002Flink + Iceberg v2 \u63d0\u4f9b\u7684\u662f\u5206\u949f\u7ea7\u522b\u7684\u89e3\u51b3\u65b9\u6848\uff0c\u7ed9\u4e0b\u6e38\u67e5\u8be2 ETL \u5e26\u6765\u4e86\u6781\u5927\u4f18\u52bf\u3002<\/p>\n<h2>\u56db\u3001\u5b9e\u65f6\u66f4\u65b0\u5c0f\u6587\u4ef6\u95ee\u9898\u7684\u4f18\u5316<\/h2>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228641\" alt=\"10\" title=\"10\" loading=\"lazy\"><\/p>\n<p>\u521a\u624d\u63d0\u5230\u7684\u5b9e\u65f6\u5c0f\u6587\u4ef6\u95ee\u9898\uff0c\u4f1a\u5728\u5f88\u5927\u7a0b\u5ea6\u4e0a\u5f71\u54cd\u4e0b\u6e38\u67e5\u8be2\u4efb\u52a1\u7684\u6027\u80fd\u3002\u63a5\u4e0b\u6765\uff0c\u7740\u91cd\u4ecb\u7ecd\u4e00\u4e0b\u6211\u4eec\u5982\u4f55\u89e3\u51b3\u5c0f\u6587\u4ef6\u95ee\u9898\u7684\u3002\u9996\u5148\uff0c\u4ecb\u7ecd\u4e00\u4e0b Iceberg Sink \u7684\u5199\u5165\u6a21\u5f0f\u3002\u7531\u4e8e\u5b58\u5728\u66f4\u65b0\u6570\u636e\u7684\u60c5\u51b5\uff0c\u6240\u4ee5\u9009\u62e9\u4f7f\u7528 Upsert Mode\u3002<\/p>\n<p>\u5728\u6bcf\u6b21\u5199\u5165\u6570\u636e\u7684\u8fc7\u7a0b\u4e2d\uff0c\u4f1a\u751f\u6210\u4e24\u6761 Record \u6570\u636e\uff0c\u5373 Delete \u548c Insert\u3002\u5728\u4e00\u5b9a\u7a0b\u5ea6\u4e0a\uff0c\u8fd9\u79cd\u65b9\u5f0f\u9020\u6210\u4e86\u5b58\u50a8\u7a7a\u95f4\u7684\u6d6a\u8d39\u3002\u4e0b\u6e38 Writer \u7b97\u5b50\u4f1a\u6709 CPU \u538b\u529b\uff0c\u5b83\u9700\u8981\u5904\u7406\u7684\u6570\u636e\u91cf\u66f4\u591a\uff0c\u9700\u8981\u5199\u5165\u7684\u6570\u636e\u66f4\u591a\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228642\" alt=\"11\" title=\"11\" loading=\"lazy\"><\/p>\n<p>\u901a\u8fc7\u5f15\u5165 Flink State \u7684\u65b9\u5f0f\uff0c\u5728\u4e00\u5b9a\u7a0b\u5ea6\u4e0a\u89e3\u51b3\u4e86 Upsert \u5199\u5165\u591a\u884c\u7684\u95ee\u9898\u3002\u9996\u5148\uff0c\u6309\u7167\u5e7f\u544a\u4e3b\u952e\u8fdb\u884c KeyBy Stream\u3002\u5982\u679c\u5f53\u524d\u4e3b\u952e\u4e0d\u5728 Flink State \u4e2d\uff0c\u8fd9\u6761\u6570\u636e\u662f\u7b2c\u4e00\u6b21\u5199\u5165\uff0c\u4f1a\u5411\u4e0b\u6e38\u8f93\u51fa\u4e00\u6761 RowKind INSERT \u6570\u636e\uff0c\u8868\u660e\u8fd9\u662f\u4e00\u6761\u5168\u65b0\u7684\u6570\u636e\u3002<\/p>\n<p>\u5982\u679c\u8be5\u6570\u636e\u4e3b\u952e\u5df2\u7ecf\u5b58\u5728\u4e8e Flink State \u4e2d\uff0c\u4f1a\u5411\u4e0b\u6e38\u8f93\u51fa\u4e24\u6761\u8bb0\u5f55\u3002\u4e00\u6761\u662f UPDATE_BEFORE\uff0c\u53e6\u4e00\u6761\u662f UPDATE_AFTER\u3002\u5728\u8fd9\u4e00\u73af\u8282\uff0c\u4f1a\u66f4\u52a0\u8be6\u7ec6\u7684\u68c0\u67e5\u662f\u5426\u9700\u8981\u8f93\u51fa\uff0c\u6bd4\u5982\u662f\u5426\u6709\u65f6\u95f4\u6233\u7684\u66f4\u65b0\uff0c\u662f\u5426\u6709\u7ef4\u8868\u4fe1\u606f\u66f4\u65b0\u7b49\u7b49\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228643\" alt=\"12\" title=\"12\" loading=\"lazy\"><\/p>\n<p>\u901a\u8fc7\u8fd9\u4e9b\u64cd\u4f5c\uff0c\u53ef\u4ee5\u5728\u4e00\u5b9a\u7a0b\u5ea6\u4e0a\uff0c\u51cf\u5c11\u4e00\u90e8\u5206\u7684\u5c0f\u6587\u4ef6\u3002\u4f46\u5728\u5b9e\u9645\u60c5\u51b5\u4e0b\uff0c\u6211\u4eec\u53d1\u73b0\u8be5\u65b9\u6cd5\u4ecd\u6709\u4e0d\u8db3\uff0c\u4f9d\u7136\u4f1a\u6709\u5f88\u591a\u7684\u5c0f\u6587\u4ef6\u751f\u6210\u3002\u57fa\u4e8e Iceberg Flink Sink \u539f\u7406\uff0c\u5927\u91cf\u7684\u5c0f\u6587\u4ef6\u901a\u8fc7 IcebergStreamWriter \u751f\u6210\u7684\u3002<\/p>\n<p>Iceberg \u652f\u6301\u4e24\u79cd\u4e0d\u540c\u7684 Distribution \u6a21\u5f0f\uff0c\u5c06\u6570\u636e\u4ece\u4e0a\u6e38\u7684 input stream\uff0c\u4f20\u8f93\u5230\u4e0b\u6e38\u7684 Writer \u7b97\u5b50\u4e2d\u3002\u7b2c\u4e00\u79cd\u662f Equality Field KeySelector\uff0c\u5373\u5c06RowData\u7684equality filed \u8fdb\u884c hash\u3002\u7b2c\u4e8c\u79cd\u662f PartitionKeySelector\uff0c\u5373\u5c06 RowData \u7684 parition field \u8fdb\u884c hash\u3002<\/p>\n<p>\u8fd9\u4e24\u79cd\u65b9\u5f0f\u6709\u4ec0\u4e48\u533a\u522b\u5462\uff1fEquality Field KeySelector \u4ece\u8bed\u4e49\u4e0a\u53ef\u4ee5\u7406\u89e3\u4e3a\u5c06 RowData \u4ee5\u4e3b\u952e hash \u7684\u65b9\u5f0f\u4f20\u8f93\u5230\u4e0b\u6e38\uff0c\u8fd9\u6837\u53ef\u4ee5\u6700\u5927\u5316\u4f7f\u7528\u4e0b\u6e38 Writer \u7b97\u5b50\u7684\u5199\u51fa\u901f\u5ea6\u3002\u800c partitionKeySelector\uff0c\u53ef\u4ee5\u5c06\u5177\u6709\u76f8\u540c Parition \u7684 RowData \u8f93\u51fa\u5230\u540c\u4e00\u4e2a Writer\uff0c\u786e\u4fdd\u540c\u4e00\u4e2a Partition \u7684\u6570\u636e\u90fd\u662f\u901a\u8fc7\u540c\u4e00\u4e2a Writer \u5199\u51fa\u3002<\/p>\n<p>StreamWriter \u8d1f\u8d23\u5c06\u6240\u6709\u6536\u5230\u7684\u6570\u636e\u8f93\u51fa\u5230 DFS\uff0c\u6bd4\u5982 S3\u4e0a\uff0c\u8fd9\u91cc\u4f1a\u6839\u636e\u8868\u4e0a\u662f\u5426\u5e26\u6709 Partition \u4fe1\u606f\u6765\u533a\u5206\u5230\u5e95\u662f\u8f93\u51fa\u5230\u540c\u4e00\u4e2a\u6587\u4ef6\u8fd8\u662f\u591a\u4e2a\u6587\u4ef6\u3002<\/p>\n<p>\u5728\u6211\u4eec\u8fd9\u4e2a\u7528\u4f8b\u4e2d\uff0c\u6570\u636e\u6e56\u662f\u6309\u7167 Partition \u6765\u8fdb\u884c\u7269\u7406\u5206\u533a\uff0c\u5373\u540c\u4e00\u4e2a\u5c0f\u65f6\u7684\u6570\u636e\u53ea\u4f1a\u5b58\u5728\u540c\u4e00\u4e2a\u8def\u5f84\u4e0b\u9762\uff0c\u800c\u540c\u4e00\u4e2a\u6570\u636e\u6587\u4ef6\u4e0d\u80fd\u5305\u542b\u591a\u4e2a Partition \u7684\u6570\u636e\u3002\u4e0b\u6e38\u7684 Writer \u5728\u6536\u5230\u6570\u636e\u4ee5\u540e\uff0c\u5c31\u4f1a\u6309\u7167 Partition \u7684\u4fe1\u606f\u6765\u5199\u51fa\u6587\u4ef6\u3002<\/p>\n<p>\u6240\u6709\u7684 Writer \u5728 Checkpoint \u9636\u6bb5\u4f1a\u5c06\u5199\u51fa\u53bb\u7684\u6587\u4ef6\u7edf\u8ba1\u4fe1\u606f\u53d1\u9001\u5230\u6700\u540e\u7684 Committer \u7b97\u5b50\u3002Commit \u7b97\u5b50\u518d\u5c06\u6240\u6709\u7684\u4fee\u6539\u63d0\u4ea4\u5230 Iceberg \u4e2d\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228644\" alt=\"13\" title=\"13\" loading=\"lazy\"><\/p>\n<p>Equality Field KeySelector \u662f\u6309\u7167 Record \u4e3b\u952e\uff0cShuffle \u5230\u4e0b\u6e38 Writer \u4e2d\u3002<\/p>\n<p>\u5728\u540c\u4e00\u4e2a Partition \u8def\u5f84\u4e0b\u9762\uff0c\u4f1a\u6709\u591a\u4e2a Writer \u540c\u65f6\u5199\u5165\u3002\u4e3b\u8981\u539f\u56e0\u5c31\u662f\u4e0b\u6e38 Writer \u63a5\u6536\u5230\u7684 RowData \u662f\u6309\u7167\u4e3b\u952e\u6765\u8fdb\u884c hash  Shuffle \u7684\uff0c\u6240\u4ee5\u6bcf\u4e2a Writer \u7b97\u5b50\u90fd\u6709\u53ef\u80fd\u63a5\u6536\u5230\u540c\u4e00\u4e2a Partition \u4e0b\u7684\u6570\u636e\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228645\" alt=\"14\" title=\"14\" loading=\"lazy\"><\/p>\n<p>\u5047\u8bbe Checkpoint \u7684\u95f4\u9694\u4e3a 20 \u5206\u949f\uff0c\u4f7f\u7528 10 \u4e2a Writer \u53bb\u5199\u6587\u4ef6\u3002\u7406\u8bba\u4e0a\uff0c\u6bcf\u4e2a\u5c0f\u65f6\u53ef\u4ee5\u5199\u51fa 90 \u4e2a\u5c0f\u6587\u4ef6\uff0c\u662f\u975e\u5e38\u7684\u5178\u578b\u7684\u957f\u5c3e\u578b\u6570\u636e\u5206\u5e03\u3002\u7531\u6b64\u53ef\u89c1\uff0c\u8d8a\u9760\u8fd1\u5f53\u524d\u5c0f\u65f6\uff0c\u9700\u8981\u5904\u7406\u7684\u6570\u636e\u91cf\u8d8a\u5927\u7684\u3002\u5982\u679c\u8ddd\u79bb\u5f53\u524d\u5c0f\u65f6\u8d8a\u8fdc\uff0c\u9700\u8981\u5904\u7406\u7684\u6570\u91cf\u975e\u5e38\u5c0f\u3002\u5bf9\u4e8e\u8fd9\u4e9b Partition \u6765\u8bf4\uff0c\u5b83\u4eec\u9700\u8981\u751f\u6210\u7684\u6587\u4ef6\u6570\u91cf\u57fa\u672c\u6052\u5b9a\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228646\" alt=\"15\" title=\"15\" loading=\"lazy\"><\/p>\n<p>PartitionKeySelector \u6309\u7167 Record \u7684 Partition \u4fe1\u606f\uff0cShuffle \u5230\u4e0b\u6e38 Writer\u3002\u5728\u540c\u4e00\u4e2a Partition \u8def\u5f84\u4e0b\uff0c\u53ea\u6709 1 \u4e2a Writer \u5199\u5165\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228647\" alt=\"16\" title=\"16\" loading=\"lazy\"><\/p>\n<p>\u5047\u8bbe Checkpoint \u7684\u95f4\u9694\u4e3a 20 \u5206\u949f\uff0c\u4f7f\u7528 1 \u4e2a Writer \u53bb\u5199\u6587\u4ef6\u3002\u8d8a\u9760\u8fd1\u6700\u65b0\u65f6\u95f4\uff0c\u5b83\u7684\u53cd\u538b\u8d8a\u4e25\u91cd\uff0c\u5bfc\u81f4\u6574\u4e2a Flink \u4f5c\u4e1a\u5ef6\u8fdf\u3002\u56e0\u4e3a\u8d8a\u662f\u9760\u8fd1\u5f53\u524d\u5c0f\u65f6\uff0c\u9700\u8981\u5904\u7406\u7684\u6570\u91cf\u7ea7\u8d8a\u5927\u3002\u8d8a\u8fdc\u79bb\u5f53\u524d\u5c0f\u65f6\uff0c\u9700\u8981\u5904\u7406\u7684\u6570\u636e\u91cf\u662f\u8d8a\u5c0f\u3002<\/p>\n<p>Equality Field KeySelector \u7684\u4f18\u52bf\u662f\u9ad8\u6548\uff0c\u4f46\u95ee\u9898\u5728\u4e8e\u5c0f\u6587\u4ef6\u7279\u522b\u591a\u3002\u5c24\u5176\u5728\u957f\u5c3e\u672b\u7aef\uff0c\u5e73\u5747\u90fd\u662f\u51e0\u5341 kb \u7684\u5c0f\u6587\u4ef6\u3002PartitionKeySelector \u7684\u4f18\u52bf\u5728\u4e8e\uff0c\u5c0f\u6587\u4ef6\u6570\u91cf\u5c11\uff0c\u5bf9\u4e8e\u6570\u636e\u91cf\u8f83\u5927\u7684 Partition\uff0c\u4f1a\u9020\u6210\u5f88\u5927\u7684\u53cd\u538b\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228648\" alt=\"17\" title=\"17\" loading=\"lazy\"><\/p>\n<p>\u4e3a\u4e86\u89e3\u51b3\u4e0a\u8ff0\u95ee\u9898\uff0c\u6211\u4eec\u5f15\u5165\u4e86 Dynamic Shuffle Operator \u7b97\u5b50\u3002\u5b83\u53ef\u4ee5\u6309\u7167\u4e0d\u540c\u7684 Partition\uff0c\u9009\u62e9\u4e0d\u540c\u7684 KeySelector\u3002<\/p>\n<p>\u6bd4\u5982\u6700\u8fd1\u7684 Partition \u6570\u636e\u91cf\u7279\u522b\u5927\uff0cDynamic Shuffle Operator \u4f1a\u9009\u62e9\u4f7f\u7528 Equality Field KeySelector\u3002\u9762\u5bf9\u957f\u5c3e\u7684 Partition\uff0cDynamic Shuffle Operator \u4f1a\u9009\u62e9 PartitionKeySelector\u3002\u8be5\u65b9\u6848\u65e2\u4fdd\u8bc1\u4e86\u5927\u6279\u91cf\u7684 Partition \u6570\u636e\uff0c\u53ef\u4ee5\u53ca\u65f6\u8f93\u51fa\u5230\u6587\u4ef6\u4e2d\uff0c\u4e5f\u51cf\u5c11\u4e86\u5728\u957f\u5c3e\u672b\u7aef\u751f\u6210\u7684\u5c0f\u6587\u4ef6\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228649\" alt=\"18\" title=\"18\" loading=\"lazy\"><\/p>\n<p>\u5728\u8fd9\u4e2a\u89e3\u51b3\u65b9\u6848\u4e2d\uff0c\u901a\u8fc7\u5f15\u5165 Dynamic Shuffle Operator\uff0c\u5728\u6570\u636e\u8f93\u5165\u5230 Writer \u524d\uff0c\u5148\u901a\u8fc7 Dynamic Shuffle Operator \u8fdb\u884c\u4e00\u6b21\u7269\u7406 Partition\uff0c\u5373\u7269\u7406\u5206\u533a\u3002<\/p>\n<p>\u800c Partition \u7b56\u7565\u4f1a\u6309\u7167 Shuffle Operator \u8fc7\u53bb\u5904\u7406\u7684\u7edf\u8ba1\u4fe1\u606f\uff0c\u8fdb\u884c\u52a8\u6001\u7f16\u6392\u3002\u5982\u4e0a\u56fe\u6240\u793a\uff0c\u9996\u5148\u901a\u8fc7\u5f15\u5165 Shuffle Coordinator \u89e3\u51b3\u4e0d\u540c Shuffle Subtask \u4e4b\u95f4\u7684\u4fe1\u606f\u901a\u4fe1\u95ee\u9898\u3002<\/p>\n<p>\u5176\u6b21\u6211\u4eec\u9700\u8981\u786e\u4fdd\u7684\u662f\u4e0d\u540c\u7684 Subtask \u5728\u8f93\u51fa\u6587\u4ef6\u7684\u65f6\u5019\u6309\u7167\u540c\u4e00\u4e2aShuffleStrategy \u6765\u8fdb\u884c\u8f93\u51fa\uff0c\u56e0\u4e3a Iceberg \u5728\u5904\u7406 Delete \u6587\u4ef6\u65f6\uff0c\u9700\u8981\u540c\u4e00\u4e2a\u4e3b\u952e\u7684 RowData \u5728\u76f8\u540c\u7684 Writer \u8f93\u51fa\uff0c\u6bd4\u5982\u6211\u4eec\u73b0\u6709\u4e00\u6761 insert\uff0c\u518d\u6765\u4e00\u6761 Update\uff0c\u5982\u679c\u8fd9\u4e24\u4e2a RowData \u662f\u6309\u7167\u4e0d\u540c\u7684 ShuffleStrategy \u6765\u8fdb\u884c Shuffle\uff0c\u5f88\u6709\u53ef\u80fd\u8fd9\u4e24\u4e2a\u6570\u636e\u4f1a Shuffle \u5230\u4e0d\u540c\u7684 Writer \u7b97\u5b50\uff0c\u8fd9\u6837\u4f1a\u5bfc\u81f4\u91cd\u590d\u6570\u636e\u7684\u4ea7\u751f\u3002<\/p>\n<p>\u9664\u6b64\u4e4b\u5916\uff0cShuffle Operator \u8d1f\u8d23\u5c06\u5df2\u7ecf\u5904\u7406\u7684\u7edf\u8ba1\u4fe1\u606f\u53d1\u9001\u7ed9 Coordinator\u3002\u6bd4\u5982\u5404\u4e2a Partition \u5904\u7406\u7684\u6570\u636e\u91cf\u3002<\/p>\n<p>\u5176\u76ee\u7684\u662f\uff0cCoordinator \u5728\u6536\u96c6\u5230 Shuffle Operator \u7684\u7edf\u8ba1\u4fe1\u606f\u4e4b\u540e\uff0c\u53ef\u4ee5\u6309\u7167\u5386\u53f2\u4fe1\u606f\u52a8\u6001\u7684\u5224\u65ad\u51fa\uff0c\u6700\u65b0\u7684 Partition \u9700\u8981\u4ec0\u4e48\u6837\u7684 Strategy\u3002\u6bd4\u5982\u5f53\u6700\u65b0\u7684 Partition \u5df2\u7ecf\u5199\u51fa\u4e86 70%\u7684\u6570\u636e\u65f6\uff0cCoordinator \u53ef\u4ee5\u8ba9 Shuffle Operator \u5207\u6362\u5230 PartitionKey\uff0c\u4ece\u800c\u51cf\u5c11\u5c0f\u6587\u4ef6\u7684\u6570\u91cf\u751f\u6210\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228650\" alt=\"19\" title=\"19\" loading=\"lazy\"><\/p>\n<p>\u7efc\u4e0a\u6240\u8ff0\uff0cDynamic Shuffle KeySelector \u6309\u7167\u5f53\u524d\u6700\u5927 PartitionKey \u6765\u5206\u914d ShuffleStrategy\uff1b\u6309\u7167\u5386\u53f2\u6570\u636e\u4fe1\u606f\u6765\u52a8\u6001\u5206\u914d Shuffle Strategy\uff0c\u6700\u7ec8\u786e\u4fdd\u6240\u6709 Subtask \u90fd\u4f7f\u7528\u76f8\u540c\u7684 Shuffle Strategy\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228651\" alt=\"20\" title=\"20\" loading=\"lazy\"><\/p>\n<p>\u63a5\u4e0b\u6765\uff0c\u4ecb\u7ecd\u4e00\u4e0b\u76f8\u5173\u7684\u5b9e\u9a8c\u5bf9\u6bd4\u3002\u6211\u4eec\u5bf9\u6bd4\u4e86 24 \u5c0f\u65f6\u4ee5\u5185\uff0c\u6bcf\u5c0f\u65f6\u6587\u4ef6\u751f\u6210\u7684\u6570\u91cf\u4ee5\u53ca\u5e73\u5747\u5927\u5c0f\u3002\u6211\u4eec\u5c06 Flink \u5e76\u53d1\u8bbe\u7f6e\u4e3a 20\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228652\" alt=\"21\" title=\"21\" loading=\"lazy\"><\/p>\n<p>\u5982\u4e0a\u8868\u6240\u793a\uff0c\u9996\u5148\u6211\u4eec\u6bd4\u8f83\u76f8\u540c Partition \u6bcf\u5c0f\u65f6\u65b0\u589e\u6587\u4ef6\u6570\u91cf\uff0c+1 \u8868\u793a\u6bd4\u6700\u65b0\u7684\u5c0f\u65f6\u665a\u4e00\u4e2a\u5c0f\u65f6\u3002No Shuffle \u8868\u793a\u662f\u7528 Iceberg \u7684\u9ed8\u8ba4 Shuffle\uff0c\u5373 EqualityFieldKeyBy\u3002Dynamicshuffle \u662f\u65b0\u7684 Shuffle Strategy\u3002<\/p>\n<p>\u53ef\u4ee5\u770b\u5230\u4e0d\u4ec5\u5728\u6700\u65b0\u7684\u51e0\u4e2a Partition \u4e2d\uff0cDynamic Shuffle \u5199\u51fa\u4e86\u66f4\u5c11\u7684\u6587\u4ef6\u6570\u91cf\uff0c\u800c\u4e14\u5728\u957f\u5c3e\u7684 Partition \u4e5f\u6709\u66f4\u597d\u7684\u6548\u679c\u3002<\/p>\n<p>\u4e00\u822c\u6765\u8bf4 \u5f53\u8fc7\u4e86 1 \u4e2a\u5c0f\u65f6\u5019\u4e4b\u540e\uff0cDynamic Shuffle Operator \u5c31\u4f1a\u5c06\u8be5 Partition \u7684 Strategy \u5207\u6362\u4e3a Partitionkeyby\uff0c\u56e0\u6b64\u5f53\u524d\u5c0f\u65f6\u7684\u6587\u4ef6\u589e\u957f\u901f\u7387\u5c31\u662f\u57fa\u672c\u6052\u5b9a\u7684\u3002<\/p>\n<p>\u53f3\u4fa7\u7684\u56fe\u4e5f\u53ef\u4ee5\u5728\u53cd\u5e94\u8fd9\u4e2a\u957f\u5c3e\u7684\u73b0\u8c61\uff1a\u53ef\u4ee5\u770b\u51fa\u6587\u4ef6\u751f\u6210\u7684\u9ad8\u5cf0\u4e00\u822c\u90fd\u662f\u5728\u7b2c\u4e00\u4e2a\u5c0f\u65f6\uff0c\u800c\u540e\u7eed\u957f\u5c3e\u5c0f\u65f6\u57fa\u672c\u662f\u56fa\u5b9a\u7684\u3002<\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228653\" alt=\"22\" title=\"22\" loading=\"lazy\"><\/p>\n<p>\u5bf9\u4e8e\u6587\u4ef6\u7684\u5e73\u5747\u5927\u5c0f\uff0cDynamic Shuffle Operator \u4e5f\u6709\u66f4\u597d\u7684\u8868\u73b0\u3002\u7531\u4e8e\u8fd9\u91cc\u91c7\u53d6\u7684\u6307\u6807\u662f\u5e73\u5747\u6587\u4ef6\u5927\u5c0f\uff0c\u800c\u4e00\u6b21 Writer \u7684\u5199\u5165\u53ef\u80fd\u4f1a\u6709\u5f88\u5927\u7684 Data \u6587\u4ef6\uff0c\u4f46 Delete \u6587\u4ef6\u901a\u5e38\u8f83\u5c0f\u3002\u56e0\u4e3a\u53ea\u5305\u542b\u4e86\u90e8\u5206\u4e3b\u952e\u6216\u8005\u4f4d\u7f6e\u4fe1\u606f\u3002\u6700\u8fd1\u5c0f\u65f6\u7684\u5e73\u5747\u5927\u5c0f\u6548\u679c\u6bd4\u8f83\u663e\u8457\u3002<\/p>\n<h2>\u4e94\u3001\u603b\u7ed3\u4e0e\u5c55\u671b<\/h2>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228654\" alt=\"23\" title=\"23\" loading=\"lazy\"><\/p>\n<p><a target=\"_blank\" href=\"https:\/\/link.segmentfault.com\/?enc=Zewt5BpzQwaPcEIv0dmsNw%3D%3D.JfoLCUAGJd9kNSq5hj3yH6AWAwEt3UoJgBi3n%2FUcMnDxABiXJqM762Ouk85cM5P3q4zzd87IAoS%2B3TZVEzucMsZK%2BU7ORYUc%2BJGmLWKsuBVUzoISRsUbfADU4Og97tWmZqGNh%2BZBtVFpMtJE03evSA%3D%3D\" rel=\"noopener noreferrer\">\u70b9\u51fb\u67e5\u770b\u539f\u6587\u89c6\u9891 &#038; \u6f14\u8bb2PPT<\/a><\/p>\n<hr>\n<h3>\u66f4\u591a\u5185\u5bb9<\/h3>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000043814703\" alt=\"\" title=\"\" loading=\"lazy\"><\/p>\n<hr>\n<p><strong>\u6d3b\u52a8\u63a8\u8350<\/strong><\/p>\n<p>\u963f\u91cc\u4e91\u57fa\u4e8e Apache Flink \u6784\u5efa\u7684\u4f01\u4e1a\u7ea7\u4ea7\u54c1-\u5b9e\u65f6\u8ba1\u7b97 Flink \u7248\u73b0\u5f00\u542f\u6d3b\u52a8\uff1a<br \/>0 \u5143\u8bd5\u7528 <a target=\"_blank\" href=\"https:\/\/link.segmentfault.com\/?enc=KRpSXOUew%2B%2FmX1EgxSHKkA%3D%3D.7ll7AacW5w81pOOlXzaNbIlwqfn%2FA3Kp1EQDH83l6GOip15KYq0IrTN10N3s63hA\" rel=\"noopener noreferrer\">\u5b9e\u65f6\u8ba1\u7b97 Flink \u7248<\/a>\uff085000CU*\u5c0f\u65f6\uff0c3 \u4e2a\u6708\u5185\uff09<br \/>\u4e86\u89e3\u6d3b\u52a8\u8be6\u60c5\uff1a<a target=\"_blank\" href=\"https:\/\/link.segmentfault.com\/?enc=JU5PmDiKPetOyzEvAMzCmQ%3D%3D.qG9WVxRn1LUlZ052vJOma1gFrae4mRQsttFU15jK7ZVOcXrCegggpuiCbYjs7%2Frn\" rel=\"noopener noreferrer\">https:\/\/free.aliyun.com\/?pipCode=sc<\/a><\/p>\n<p><img decoding=\"async\"   referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000044228655\" alt=\"image.png\" title=\"image.png\" loading=\"lazy\"><\/p>\n<p><a href=\"https:\/\/segmentfault.com\/a\/1190000044228630\">https:\/\/segmentfault.com\/a\/1190000044228630<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u6458\u8981\uff1a\u672c\u6587\u6574\u7406\u81ea SmartNews \u6570\u636e\u5e73\u53f0\u67b6\u6784\u5e08 Apache Iceberg Contributor \u6222 [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[4],"tags":[],"class_list":["post-10008274","post","type-post","status-publish","format-standard","hentry","category-share"],"jetpack_sharing_enabled":true,"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/itmax.cn\/index.php?rest_route=\/wp\/v2\/posts\/10008274","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/itmax.cn\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/itmax.cn\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/itmax.cn\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/itmax.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=10008274"}],"version-history":[{"count":0,"href":"https:\/\/itmax.cn\/index.php?rest_route=\/wp\/v2\/posts\/10008274\/revisions"}],"wp:attachment":[{"href":"https:\/\/itmax.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=10008274"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/itmax.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=10008274"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/itmax.cn\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=10008274"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}