Closed wangkuiyi closed 4 years ago
I Googled around and noticed that all results I found use INSERT INTO TABLE
.
@typhoonzero Should we change FeatureDerivationCaseSQLHive? Or, we don't have to because of HiveQL tolerant with the lack of keyword TABLE?
I think it's better to keep all Hive SQL's all use INSERT INTO TABLE
.
Done in #1595
I noticed that the FeatureDerivationCaseSQLHive SQL program uses
INSERT INTO
other thanINSERT INTO TABLE
.https://github.com/sql-machine-learning/sqlflow/blob/f3c631e27e520310e1817300799967589a908c32/pkg/sql/testdata/feature_derivation_case.go#L45
Whereas other HiveQL programs use
INSERT INTO TABLE
. For examplehttps://github.com/sql-machine-learning/sqlflow/blob/f3c631e27e520310e1817300799967589a908c32/pkg/sql/testdata/iris_sql.go#L187
https://github.com/sql-machine-learning/sqlflow/blob/f3c631e27e520310e1817300799967589a908c32/pkg/sql/testdata/iris_sql.go#L306
https://github.com/sql-machine-learning/sqlflow/blob/f3c631e27e520310e1817300799967589a908c32/pkg/sql/testdata/churn_sql.go#L188
I noticed that https://github.com/sql-machine-learning/sqlflow/pull/1247/files#diff-55c6e51a3df3139780b8d2a289b85a03R45 adds FeatureDerivationCaseSQLHive.
@typhoonzero Should we change FeatureDerivationCaseSQLHive? Or, we don't have to because of HiveQL tolerant with the lack of keyword
TABLE
?