[
https://issues.apache.org/jira/browse/DRILL-4970?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16069379#comment-16069379
]
ASF GitHub Bot commented on DRILL-4970:
---------------------------------------
Github user paul-rogers commented on a diff in the pull request:
https://github.com/apache/drill/pull/863#discussion_r124954136
--- Diff:
exec/java-exec/src/test/java/org/apache/drill/exec/fn/impl/TestCastFunctions.java
---
@@ -78,4 +81,71 @@ public void testToDateForTimeStamp() throws Exception {
.build()
.run();
}
+
+ @Test // DRILL-4970
+ public void testCastNegativeFloatToInt() throws Exception {
+ try {
+ test("create table dfs_test.tmp.table_with_float as\n" +
+ "(select cast(-255.0 as double) as double_col,\n" +
+ "cast(-255.0 as float) as float_col\n" +
+ "from (values(1)))");
+
+ final List<String> columnNames = Lists.newArrayList();
+ columnNames.add("float_col");
+ columnNames.add("double_col");
+
+ final List<String> castTypes = Lists.newArrayList();
+ castTypes.add("int");
+ castTypes.add("bigInt");
+
+ final String query = "select count(*) as c from
dfs_test.tmp.table_with_float\n" +
+ "where (cast(%1$s as %2$s) >= -255 and (%1$s
<= -5)) or (%1$s <= -256)";
+
+ for (String columnName : columnNames) {
+ for (String castType : castTypes) {
+ testBuilder()
+ .sqlQuery(query, columnName, castType)
+ .unOrdered()
+ .baselineColumns("c")
+ .baselineValues(1L)
+ .build()
+ .run();
+ }
+ }
+ } finally {
+ test("drop table if exists dfs_test.tmp.table_with_float");
+ }
+ }
+
+ @Test // DRILL-4970
+ public void testCastNegativeDecimalToVarChar() throws Exception {
--- End diff --
There are two tests here. Should we not test all cast functions to
determine if others have similar issues: (short, int, long, float, double,
decimal 9, decimal 18, etc.) --> (short, int, long, float, double, decimal 9,
decimal 18, etc.)
This is a reason to do basic JUnit tests without the server: much easier to
try all the boundary values for every type to every other type. That testing is
tedious, yes. But, much less tedious using the functions directly.
> Wrong results when casting double to bigint or int
> --------------------------------------------------
>
> Key: DRILL-4970
> URL: https://issues.apache.org/jira/browse/DRILL-4970
> Project: Apache Drill
> Issue Type: Bug
> Components: Execution - Data Types
> Affects Versions: 1.8.0
> Reporter: Robert Hou
> Assignee: Volodymyr Vysotskyi
> Attachments: test_table
>
>
> This query returns the wrong result
> {code:sql}
> 0: jdbc:drill:zk=10.10.100.186:5181/drill/rho> select count(\*) from
> test_table where (int_id > -3025 and bigint_id <= -256) or (cast(double_id as
> bigint) >= -255 and double_id <= -5);
> +---------+
> | EXPR$0 |
> +---------+
> | 2769 |
> +---------+
> {code}
> Without the cast, it returns the correct result:
> {code:sql}
> 0: jdbc:drill:zk=10.10.100.186:5181/drill/rho> select count(\*) from
> test_table where (int_id > -3025 and bigint_id <= -256) or (double_id >= -255
> and double_id <= -5);
> +---------+
> | EXPR$0 |
> +---------+
> | 3020 |
> +---------+
> {code}
> By itself, the result is also correct:
> {code:sql}
> 0: jdbc:drill:zk=10.10.100.186:5181/drill/rho> select count(\*) from
> test_table where (cast(double_id as bigint) >= -255 and double_id <= -5);
> +---------+
> | EXPR$0 |
> +---------+
> | 251 |
> +---------+
> {code}
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
