SELECT DOUBLE(1.0/2.0) FROM EXAMPLE; On Tue, Apr 28, 2020 at 2:36 PM Kwanghee Park <beb...@gmail.com> wrote:
> SELECT CAST(1.0 AS DOUBLE)/2.0 FROM EXAMPLE; > SELECT DOUBLE(1.0)/2.0 FROM EXAMPLE; > > On Tue, Apr 28, 2020 at 5:31 AM Arup Malakar <amala...@gmail.com> wrote: > >> >> Would anyone know if it could be related to calcite versions and any >> possible workarounds? >> >> On Thu, Apr 23, 2020 at 6:16 PM Arup Malakar <amala...@gmail.com> wrote: >> >>> Hi Hive Users, >>> >>> I noticed that with cbo on the query: "SELECT 1.0/2.0 FROM example limit >>> 1" returns 0 >>> >>> hive> SET hive.cbo.enable=true; SELECT 1.0/2.0 FROM example limit 1; >>>> 0 >>>> >>> >>> >>> With cbo disabled it returns 0.5 as expected: >>> >>> hive> SET hive.cbo.enable=false; SELECT 1.0/2.0 FROM example limit 1; >>>> 0.5 >>>> >>> >>> Not referring to a table returns 0.5 as expected as well: >>> >>> hive> select 1.0/2.0; >>>> 0.5 >>>> >>> >>> >>> This is how the explain plan looks like: >>> >>> hive> SET hive.cbo.enable=true; EXPLAIN extended SELECT 1.0/2.0 FROM >>>> example limit 1; >>>> OK >>>> STAGE DEPENDENCIES: >>>> Stage-0 is a root stage >>>> STAGE PLANS: >>>> Stage: Stage-0 >>>> Fetch Operator >>>> limit: 1 >>>> Processor Tree: >>>> TableScan >>>> alias: example >>>> GatherStats: false >>>> Select Operator >>>> expressions: 0 (type: int) >>>> outputColumnNames: _col0 >>>> Limit >>>> Number of rows: 1 >>>> ListSink >>> >>> >>> I have tested that it is not an issue in 2.1.1 and 3.0, but only in >>> 2.3.6. I couldn't find a corresponding JIRA that fixes it though. Does >>> anyone know if there is a JIRA that fixes this (so that I can backport) or >>> any other suggestion? >>> >>> -- >>> Arup Malakar >>> >> >> >> -- >> Arup Malakar >> >
