Re: sql语句中当in的列表超过一定数据量的时候性能突然变差，如何解决？

2017-08-08 Thread wangzy24
Found by looking at the source code by kylin.query.scan.fuzzykey.max attribute control, the default is 200, when conditions combined with more than 200, it will range scan, if less than this value will be according to the rowkey to obtain, in kylin. If you want to modify the properties of adding

Re: sql语句中当in的列表超过一定数据量的时候性能突然变差，如何解决？

2017-08-04 Thread hongbin ma
take a look at kylin.storage.hbase.max-fuzzykey-scan in org.apache.kylin.common.KylinConfigBase#getQueryScanFuzzyKeyMax 2017-08-03 20:25 GMT+08:00 ShaoFeng Shi : > It might be related with storage layer cache. You can make more tests to > see the differences. Besides, if

Re: sql语句中当in的列表超过一定数据量的时候性能突然变差，如何解决？

2017-08-03 Thread ShaoFeng Shi
It might be related with storage layer cache. You can make more tests to see the differences. Besides, if you can provide more logs when executing these two queries, that would be good for analysis. 在 2017年8月3日 下午4:26，wangzy24 写道： > 如下两个sql，差别是第二个hotel_code