[jira] [Created] (CALCITE-1722) Lose of double precision on Filters.

["slim bouguerra (JIRA)" <ji...@apache.org>]

slim bouguerra created CALCITE-1722:
---------------------------------------

             Summary: Lose of double precision on Filters.
                 Key: CALCITE-1722
                 URL: https://issues.apache.org/jira/browse/CALCITE-1722
             Project: Calcite
          Issue Type: Bug
    Affects Versions: 1.12.0
            Reporter: slim bouguerra
            Assignee: Julian Hyde


We have 2 issue here:
1 Major, it is the lose of precision when converting filters with double values eg {code} value < 1.455 {code}
For instance this is an example of wrong plan due to this bug.
{code} 
 explain select d_year from ssb_druid where cast(d_year as double) < 1.2221;
OK
Plan optimized by CBO.

Stage-0
  Fetch Operator
    limit:-1
    Select Operator [SEL_1]
      Output:["_col0"]
      TableScan [TS_0]
        Output:["d_year"],properties:{"druid.query.json":"{\"queryType\":\"select\",\"dataSource\":\"ssb_druid\",\"descending\":false,\"intervals\":[\"1900-01-01T00:00:00.000/3000-01-01T00:00:00.000\"],\"filter\":{\"type\":\"bound\",\"dimension\":\"d_year\",\"upper\":\"12221\",\"upperStrict\":true,\"alphaNumeric\":true},\"dimensions\":[\"d_year\"],\"metrics\":[],\"granularity\":\"all\",\"pagingSpec\":{\"threshold\":16384,\"fromNext\":true},\"context\":{\"druid.query.fetch\":false}}","druid.query.type":"select"}
{code}

The second issue (Minor) is to use {code}ordering{code} instead of Alphanumeric, this is needed when comparing floats/doubles.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)