You are viewing a plain text version of this content. The canonical link for it is here.
Posted to github@beam.apache.org by GitBox <gi...@apache.org> on 2020/07/13 22:33:04 UTC
[GitHub] [beam] robinyqiu commented on a change in pull request #12174: [BEAM-10239] Support ZetaSQL NUMERIC type in BeamSQL
robinyqiu commented on a change in pull request #12174:
URL: https://github.com/apache/beam/pull/12174#discussion_r453958885
##########
File path: sdks/java/extensions/sql/zetasql/src/test/java/org/apache/beam/sdk/extensions/sql/zetasql/TestInput.java
##########
@@ -258,6 +259,14 @@
.addRows(LocalTime.of(15, 30, 0), "s")
.addRows(LocalTime.of(23, 35, 59), "s");
+ private static final Schema TABLE_WTH_NUMERIC_SCHEMA =
+ Schema.builder().addDecimalField("numeric_field").addStringField("str_field").build();
+ public static final TestBoundedTable TABLE_WITH_NUMERIC =
+ TestBoundedTable.of(TABLE_WTH_NUMERIC_SCHEMA)
+ .addRows(new BigDecimal("123.4567").setScale(9), "str1")
Review comment:
`new BigDecimal("some string").setScale(9)` is used a lot of times in this file and the test file. I would create a utility function for better style.
How about we create a new file `ZetaSqlTypesUtils.java` in /main and make a `BigDecimal bigDecimalAsNumeric(String s)`?
##########
File path: sdks/java/extensions/sql/zetasql/src/main/java/org/apache/beam/sdk/extensions/sql/zetasql/translation/ExpressionConverter.java
##########
@@ -805,6 +807,22 @@ private RexNode convertSimpleValueToRexNode(TypeKind kind, Value value) {
.makeLiteral(
value.getStringValue(), typeFactory().createSqlType(SqlTypeName.VARCHAR), true);
break;
+ case TYPE_NUMERIC:
+ // As ZetaSQL NUMERIC literal would be unparsed as DOUBLE through Calcite makeExactLiteral
Review comment:
Technically speaking, it is not "unparsed as DOUBLE". Also, I would update the comment to be more descriptive, like:
Cannot simply call makeExactLiteral() because later it will be unparsed to the string representation of the BigDecimal itself (e.g. "SELECT NUMERIC '0'" will be unparsed to "SELECT 0E-9"), and Calcite does not allow customize unparsing of SqlNumericLiteral. So we create a wrapper function here such that we can later recognize it and customize its unparsing in BeamBigQuerySqlDialect.
##########
File path: sdks/java/extensions/sql/zetasql/src/test/java/org/apache/beam/sdk/extensions/sql/zetasql/ZetaSqlDialectSpecTest.java
##########
@@ -2174,6 +2175,246 @@ public void testZetaSQLNestedQueryFive() {
pipeline.run().waitUntilFinish(Duration.standardMinutes(PIPELINE_EXECUTION_WAITTIME_MINUTES));
}
+ /////////////////////////////////////////////////////////////////////////////
+ // NUMERIC type tests
+ /////////////////////////////////////////////////////////////////////////////
+
+ @Test
+ public void testNumericLiteral() {
+ String sql =
+ "SELECT NUMERIC '0', "
+ + "NUMERIC '123456', "
+ + "NUMERIC '-3.14', "
+ + "NUMERIC '-0.54321', "
+ + "NUMERIC '1.23456e05', "
+ + "NUMERIC '-9.876e-3', "
+ + "NUMERIC '-99999999999999999999999999999.999999999', "
Review comment:
Add comment on these are the min/max values that can be represented as ZetaSQL NUMERIC?
##########
File path: sdks/java/extensions/sql/zetasql/src/test/java/org/apache/beam/sdk/extensions/sql/zetasql/ZetaSqlDialectSpecTest.java
##########
@@ -2174,6 +2175,246 @@ public void testZetaSQLNestedQueryFive() {
pipeline.run().waitUntilFinish(Duration.standardMinutes(PIPELINE_EXECUTION_WAITTIME_MINUTES));
}
+ /////////////////////////////////////////////////////////////////////////////
+ // NUMERIC type tests
+ /////////////////////////////////////////////////////////////////////////////
+
+ @Test
+ public void testNumericLiteral() {
+ String sql =
+ "SELECT NUMERIC '0', "
+ + "NUMERIC '123456', "
+ + "NUMERIC '-3.14', "
+ + "NUMERIC '-0.54321', "
+ + "NUMERIC '1.23456e05', "
+ + "NUMERIC '-9.876e-3', "
+ + "NUMERIC '-99999999999999999999999999999.999999999', "
+ + "NUMERIC '99999999999999999999999999999.999999999'";
+ ;
+
+ ZetaSQLQueryPlanner zetaSQLQueryPlanner = new ZetaSQLQueryPlanner(config);
+ BeamRelNode beamRelNode = zetaSQLQueryPlanner.convertToBeamRel(sql);
+ PCollection<Row> stream = BeamSqlRelUtils.toPCollection(pipeline, beamRelNode);
+
+ PAssert.that(stream)
+ .containsInAnyOrder(
+ Row.withSchema(
+ Schema.builder()
+ .addDecimalField("f_numeric1")
+ .addDecimalField("f_numeric2")
+ .addDecimalField("f_numeric3")
+ .addDecimalField("f_numeric4")
+ .addDecimalField("f_numeric5")
+ .addDecimalField("f_numeric6")
+ .addDecimalField("f_numeric7")
+ .addDecimalField("f_numeric8")
+ .build())
+ .addValues(new BigDecimal("0").setScale(9))
+ .addValues(new BigDecimal("123456").setScale(9))
+ .addValues(new BigDecimal("-3.14").setScale(9))
+ .addValues(new BigDecimal("-0.54321").setScale(9))
+ .addValues(new BigDecimal("123456").setScale(9))
+ .addValues(new BigDecimal("-0.009876").setScale(9))
+ .addValues(new BigDecimal("-99999999999999999999999999999.999999999"))
+ .addValues(new BigDecimal("99999999999999999999999999999.999999999"))
+ .build());
+ pipeline.run().waitUntilFinish(Duration.standardMinutes(PIPELINE_EXECUTION_WAITTIME_MINUTES));
+ }
+
+ @Test
+ public void testUnaryMinusNumeric() {
+ String sql = "SELECT - NUMERIC '1.23456e05'";
+
+ ZetaSQLQueryPlanner zetaSQLQueryPlanner = new ZetaSQLQueryPlanner(config);
+ BeamRelNode beamRelNode = zetaSQLQueryPlanner.convertToBeamRel(sql);
+ PCollection<Row> stream = BeamSqlRelUtils.toPCollection(pipeline, beamRelNode);
+
+ PAssert.that(stream)
+ .containsInAnyOrder(
+ Row.withSchema(Schema.builder().addDecimalField("f_numeric").build())
+ .addValues(new BigDecimal("-123456").setScale(9))
+ .build());
+ pipeline.run().waitUntilFinish(Duration.standardMinutes(PIPELINE_EXECUTION_WAITTIME_MINUTES));
+ }
+
+ @Test
+ public void testAddNumeric() {
+ String sql = "SELECT NUMERIC '1.23456e05' + NUMERIC '9.876e-3'";
+
+ ZetaSQLQueryPlanner zetaSQLQueryPlanner = new ZetaSQLQueryPlanner(config);
+ BeamRelNode beamRelNode = zetaSQLQueryPlanner.convertToBeamRel(sql);
+ PCollection<Row> stream = BeamSqlRelUtils.toPCollection(pipeline, beamRelNode);
+
+ PAssert.that(stream)
+ .containsInAnyOrder(
+ Row.withSchema(Schema.builder().addDecimalField("f_numeric").build())
+ .addValues(new BigDecimal("123456.009876").setScale(9))
+ .build());
+ pipeline.run().waitUntilFinish(Duration.standardMinutes(PIPELINE_EXECUTION_WAITTIME_MINUTES));
+ }
+
+ @Test
+ public void testSubNumeric() {
+ String sql = "SELECT NUMERIC '1.23456e05' - NUMERIC '-9.876e-3'";
+
+ ZetaSQLQueryPlanner zetaSQLQueryPlanner = new ZetaSQLQueryPlanner(config);
+ BeamRelNode beamRelNode = zetaSQLQueryPlanner.convertToBeamRel(sql);
+ PCollection<Row> stream = BeamSqlRelUtils.toPCollection(pipeline, beamRelNode);
+
+ PAssert.that(stream)
+ .containsInAnyOrder(
+ Row.withSchema(Schema.builder().addDecimalField("f_numeric").build())
+ .addValues(new BigDecimal("123456.009876").setScale(9))
+ .build());
+ pipeline.run().waitUntilFinish(Duration.standardMinutes(PIPELINE_EXECUTION_WAITTIME_MINUTES));
+ }
+
+ @Test
+ public void testMultiNumeric() {
+ String sql = "SELECT NUMERIC '1.23e02' * NUMERIC '-1.001e-3'";
+
+ ZetaSQLQueryPlanner zetaSQLQueryPlanner = new ZetaSQLQueryPlanner(config);
+ BeamRelNode beamRelNode = zetaSQLQueryPlanner.convertToBeamRel(sql);
+ PCollection<Row> stream = BeamSqlRelUtils.toPCollection(pipeline, beamRelNode);
+
+ PAssert.that(stream)
+ .containsInAnyOrder(
+ Row.withSchema(Schema.builder().addDecimalField("f_numeric").build())
+ .addValues(new BigDecimal("-0.123123").setScale(9))
+ .build());
+ pipeline.run().waitUntilFinish(Duration.standardMinutes(PIPELINE_EXECUTION_WAITTIME_MINUTES));
+ }
+
+ @Test
+ public void testDivNumeric() {
+ String sql = "SELECT NUMERIC '-1.23123e-1' / NUMERIC '-1.001e-3'";
+
+ ZetaSQLQueryPlanner zetaSQLQueryPlanner = new ZetaSQLQueryPlanner(config);
+ BeamRelNode beamRelNode = zetaSQLQueryPlanner.convertToBeamRel(sql);
+ PCollection<Row> stream = BeamSqlRelUtils.toPCollection(pipeline, beamRelNode);
+
+ PAssert.that(stream)
+ .containsInAnyOrder(
+ Row.withSchema(Schema.builder().addDecimalField("f_numeric").build())
+ .addValues(new BigDecimal("123").setScale(9))
+ .build());
+ pipeline.run().waitUntilFinish(Duration.standardMinutes(PIPELINE_EXECUTION_WAITTIME_MINUTES));
+ }
+
+ @Test
+ public void testModNumeric() {
+ String sql = "SELECT MOD(NUMERIC '1.23456e05', NUMERIC '5')";
+
+ ZetaSQLQueryPlanner zetaSQLQueryPlanner = new ZetaSQLQueryPlanner(config);
+ BeamRelNode beamRelNode = zetaSQLQueryPlanner.convertToBeamRel(sql);
+ PCollection<Row> stream = BeamSqlRelUtils.toPCollection(pipeline, beamRelNode);
+
+ PAssert.that(stream)
+ .containsInAnyOrder(
+ Row.withSchema(Schema.builder().addDecimalField("f_numeric").build())
+ .addValues(new BigDecimal("1").setScale(9))
+ .build());
+ pipeline.run().waitUntilFinish(Duration.standardMinutes(PIPELINE_EXECUTION_WAITTIME_MINUTES));
+ }
+
+ @Test
+ public void testFloorNumeric() {
+ String sql = "SELECT FLOOR(NUMERIC '1.23456e04'), FLOOR(NUMERIC '-1.23456e04')";
+
+ ZetaSQLQueryPlanner zetaSQLQueryPlanner = new ZetaSQLQueryPlanner(config);
+ BeamRelNode beamRelNode = zetaSQLQueryPlanner.convertToBeamRel(sql);
+ PCollection<Row> stream = BeamSqlRelUtils.toPCollection(pipeline, beamRelNode);
+
+ PAssert.that(stream)
+ .containsInAnyOrder(
+ Row.withSchema(
+ Schema.builder()
+ .addDecimalField("f_numeric1")
+ .addDecimalField("f_numeric2")
+ .build())
+ .addValues(new BigDecimal("12345").setScale(9))
+ .addValues(new BigDecimal("-12346").setScale(9))
+ .build());
+ pipeline.run().waitUntilFinish(Duration.standardMinutes(PIPELINE_EXECUTION_WAITTIME_MINUTES));
+ }
+
+ @Test
+ public void testCeilNumeric() {
+ String sql = "SELECT CEIL(NUMERIC '1.23456e04'), CEIL(NUMERIC '-1.23456e04')";
+
+ ZetaSQLQueryPlanner zetaSQLQueryPlanner = new ZetaSQLQueryPlanner(config);
+ BeamRelNode beamRelNode = zetaSQLQueryPlanner.convertToBeamRel(sql);
+ PCollection<Row> stream = BeamSqlRelUtils.toPCollection(pipeline, beamRelNode);
+
+ PAssert.that(stream)
+ .containsInAnyOrder(
+ Row.withSchema(
+ Schema.builder()
+ .addDecimalField("f_numeric1")
+ .addDecimalField("f_numeric2")
+ .build())
+ .addValues(new BigDecimal("12346").setScale(9))
+ .addValues(new BigDecimal("-12345").setScale(9))
+ .build());
+ pipeline.run().waitUntilFinish(Duration.standardMinutes(PIPELINE_EXECUTION_WAITTIME_MINUTES));
+ }
+
+ @Test
+ public void testNumericColumn() {
Review comment:
Nit: move this below `testNumericLiteral()` as we did in other places?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org