ClickHouse/tests/queries/0_stateless/02500_remove_redundant_distinct_analyzer.reference

-- Disabled query_plan_remove_redundant_distinct
Expression (Project names)
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression ((Projection + (Change column names to column identifiers + Project names)))
        Distinct (DISTINCT)
          Distinct (Preliminary DISTINCT)
            Expression ((Projection + (Change column names to column identifiers + Project names)))
              Distinct (DISTINCT)
                Distinct (Preliminary DISTINCT)
                  Expression ((Projection + Change column names to column identifiers))
                    ReadFromStorage (SystemNumbers)
-- Enabled query_plan_remove_redundant_distinct
-- DISTINCT is only in most inner subquery
-- query
SELECT DISTINCT *
FROM
(
    SELECT DISTINCT *
    FROM
    (
        SELECT DISTINCT *
        FROM numbers(3)
    )
)
-- explain
Expression ((Project names + (Projection + (Change column names to column identifiers + (Project names + (Projection + (Change column names to column identifiers + Project names)))))))
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression ((Projection + Change column names to column identifiers))
        ReadFromStorage (SystemNumbers)
-- execute
0
1
2
-- do _not_ remove DISTINCT after UNION
-- query
SELECT DISTINCT number FROM
(
    (SELECT DISTINCT number FROM numbers(1))
    UNION ALL
    (SELECT DISTINCT number FROM numbers(2))
)
ORDER BY number
-- explain
Expression (Project names)
  Distinct (DISTINCT)
    Sorting (Sorting for ORDER BY)
      Expression (Before ORDER BY)
        Distinct (Preliminary DISTINCT)
          Union
            Expression ((Projection + (Change column names to column identifiers + Project names)))
              Distinct (DISTINCT)
                Distinct (Preliminary DISTINCT)
                  Expression ((Projection + Change column names to column identifiers))
                    ReadFromStorage (SystemNumbers)
            Expression (( + ( + Project names)))
              Distinct (DISTINCT)
                Distinct (Preliminary DISTINCT)
                  Expression ((Projection + Change column names to column identifiers))
                    ReadFromStorage (SystemNumbers)
-- execute
0
1
-- do _not_ remove DISTINCT after JOIN
-- query
SELECT DISTINCT *
FROM
(
    SELECT DISTINCT number AS n
    FROM numbers(2)
) as x,
(
    SELECT DISTINCT number AS n
    FROM numbers(2)
) as y
-- explain
Expression (Project names)
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression ((Projection + DROP unused columns after JOIN))
        Join (JOIN FillRightFirst)
          Expression ((Change column names to column identifiers + Project names))
            Distinct (DISTINCT)
              Distinct (Preliminary DISTINCT)
                Expression ((Projection + Change column names to column identifiers))
                  ReadFromStorage (SystemNumbers)
          Expression ((Change column names to column identifiers + Project names))
            Distinct (DISTINCT)
              Distinct (Preliminary DISTINCT)
                Expression ((Projection + Change column names to column identifiers))
                  ReadFromStorage (SystemNumbers)
-- execute
0	0
0	1
1	0
1	1
-- DISTINCT duplicates with several columns
-- query
SELECT DISTINCT *
FROM
(
    SELECT DISTINCT *
    FROM
    (
        SELECT DISTINCT number as a, 2*number as b
        FROM numbers(3)
    )
)
-- explain
Expression ((Project names + (Projection + (Change column names to column identifiers + (Project names + (Projection + (Change column names to column identifiers + Project names)))))))
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression ((Projection + Change column names to column identifiers))
        ReadFromStorage (SystemNumbers)
-- execute
0	0
1	2
2	4
-- DISTINCT duplicates with constant columns
-- query
SELECT DISTINCT 2, a, b
FROM
(
    SELECT DISTINCT a, b
    FROM
    (
        SELECT DISTINCT 1, number as a, 2*number as b
        FROM numbers(3)
    )
)
-- explain
Expression ((Project names + (Projection + (Change column names to column identifiers + (Project names + (Projection + (Change column names to column identifiers + Project names)))))))
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression ((Projection + Change column names to column identifiers))
        ReadFromStorage (SystemNumbers)
-- execute
2	0	0
2	1	2
2	2	4
-- ARRAY JOIN: do _not_ remove outer DISTINCT because new rows are generated between inner and outer DISTINCTs
-- query
SELECT DISTINCT *
FROM
(
    SELECT DISTINCT *
    FROM VALUES('Hello', 'World', 'Goodbye')
) AS words
ARRAY JOIN [0, 1] AS arr
-- explain
Expression (Project names)
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression (Projection)
        ArrayJoin (ARRAY JOIN)
          Expression ((DROP unused columns before ARRAY JOIN + (ARRAY JOIN actions + (Change column names to column identifiers + Project names))))
            Distinct (DISTINCT)
              Distinct (Preliminary DISTINCT)
                Expression ((Projection + Change column names to column identifiers))
                  ReadFromStorage (Values)
-- execute
Hello
World
Goodbye
-- WITH FILL: do _not_ remove outer DISTINCT because new rows are generated between inner and outer DISTINCTs
-- query
SELECT DISTINCT *
FROM
(
    SELECT DISTINCT *
    FROM values('id UInt8', 0, 2)
    ORDER BY id ASC WITH FILL
)
-- explain
Expression (Project names)
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression ((Projection + (Change column names to column identifiers + Project names)))
        Filling
          Distinct (DISTINCT)
            Sorting (Sorting for ORDER BY)
              Expression (Before ORDER BY)
                Distinct (Preliminary DISTINCT)
                  Expression ((Projection + Change column names to column identifiers))
                    ReadFromStorage (Values)
-- execute
0
1
2
-- WHERE with arrayJoin(): do _not_ remove outer DISTINCT because new rows are generated between inner and outer DISTINCTs
-- query
SELECT DISTINCT *
FROM
(
    SELECT DISTINCT ['Istanbul', 'Berlin', 'Bensheim'] AS cities
)
WHERE arrayJoin(cities) IN ['Berlin', 'Bensheim']
-- explain
Expression (Project names)
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression (Projection)
        Filter ((WHERE + (Change column names to column identifiers + Project names)))
          Distinct (DISTINCT)
            Distinct (Preliminary DISTINCT)
              Expression ((Projection + Change column names to column identifiers))
                ReadFromStorage (SystemOne)
-- execute
['Istanbul','Berlin','Bensheim']
-- GROUP BY before DISTINCT with on the same columns => remove DISTINCT
-- query
SELECT DISTINCT a
FROM
(
    SELECT
        a,
        sum(b) AS c
    FROM
    (
        SELECT
            x.number AS a,
            y.number AS b
        FROM numbers(3) AS x, numbers(3, 3) AS y
    )
    GROUP BY a
)
-- explain
Expression ((Project names + (Projection + (Change column names to column identifiers + (Project names + Projection)))))
  Aggregating
    Expression ((Before GROUP BY + (Change column names to column identifiers + (Project names + (Projection + DROP unused columns after JOIN)))))
      Join (JOIN FillRightFirst)
        Expression (Change column names to column identifiers)
          ReadFromStorage (SystemNumbers)
        Expression (Change column names to column identifiers)
          ReadFromStorage (SystemNumbers)
-- execute
0
2
1
-- GROUP BY before DISTINCT with on different columns => do _not_ remove DISTINCT
-- query
SELECT DISTINCT c
FROM
(
    SELECT
        a,
        sum(b) AS c
    FROM
    (
        SELECT
            x.number AS a,
            y.number AS b
        FROM numbers(3) AS x, numbers(3, 3) AS y
    )
    GROUP BY a
)
-- explain
Expression (Project names)
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression ((Projection + (Change column names to column identifiers + (Project names + Projection))))
        Aggregating
          Expression ((Before GROUP BY + (Change column names to column identifiers + (Project names + (Projection + DROP unused columns after JOIN)))))
            Join (JOIN FillRightFirst)
              Expression (Change column names to column identifiers)
                ReadFromStorage (SystemNumbers)
              Expression (Change column names to column identifiers)
                ReadFromStorage (SystemNumbers)
-- execute
12
-- GROUP BY WITH ROLLUP before DISTINCT with on different columns => do _not_ remove DISTINCT
-- query
SELECT DISTINCT c
FROM
(
    SELECT
        a,
        sum(b) AS c
    FROM
    (
        SELECT
            x.number AS a,
            y.number AS b
        FROM numbers(3) AS x, numbers(3, 3) AS y
    )
    GROUP BY a WITH ROLLUP
)
-- explain
Expression (Project names)
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression ((Projection + (Change column names to column identifiers + (Project names + Projection))))
        Rollup
          Aggregating
            Expression ((Before GROUP BY + (Change column names to column identifiers + (Project names + (Projection + DROP unused columns after JOIN)))))
              Join (JOIN FillRightFirst)
                Expression (Change column names to column identifiers)
                  ReadFromStorage (SystemNumbers)
                Expression (Change column names to column identifiers)
                  ReadFromStorage (SystemNumbers)
-- execute
12
36
-- GROUP BY WITH ROLLUP before DISTINCT with on the same columns => remove DISTINCT
-- query
SELECT DISTINCT a
FROM
(
    SELECT
        a,
        sum(b) AS c
    FROM
    (
        SELECT
            x.number AS a,
            y.number AS b
        FROM numbers(3) AS x, numbers(3, 3) AS y
    )
    GROUP BY a WITH ROLLUP
)
-- explain
Expression ((Project names + (Projection + (Change column names to column identifiers + (Project names + Projection)))))
  Rollup
    Aggregating
      Expression ((Before GROUP BY + (Change column names to column identifiers + (Project names + (Projection + DROP unused columns after JOIN)))))
        Join (JOIN FillRightFirst)
          Expression (Change column names to column identifiers)
            ReadFromStorage (SystemNumbers)
          Expression (Change column names to column identifiers)
            ReadFromStorage (SystemNumbers)
-- execute
0
2
1
0
-- GROUP BY WITH CUBE before DISTINCT with on different columns => do _not_ remove DISTINCT
-- query
SELECT DISTINCT c
FROM
(
    SELECT
        a,
        sum(b) AS c
    FROM
    (
        SELECT
            x.number AS a,
            y.number AS b
        FROM numbers(3) AS x, numbers(3, 3) AS y
    )
    GROUP BY a WITH CUBE
)
-- explain
Expression (Project names)
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression ((Projection + (Change column names to column identifiers + (Project names + Projection))))
        Cube
          Aggregating
            Expression ((Before GROUP BY + (Change column names to column identifiers + (Project names + (Projection + DROP unused columns after JOIN)))))
              Join (JOIN FillRightFirst)
                Expression (Change column names to column identifiers)
                  ReadFromStorage (SystemNumbers)
                Expression (Change column names to column identifiers)
                  ReadFromStorage (SystemNumbers)
-- execute
12
36
-- GROUP BY WITH CUBE before DISTINCT with on the same columns => remove DISTINCT
-- query
SELECT DISTINCT a
FROM
(
    SELECT
        a,
        sum(b) AS c
    FROM
    (
        SELECT
            x.number AS a,
            y.number AS b
        FROM numbers(3) AS x, numbers(3, 3) AS y
    )
    GROUP BY a WITH CUBE
)
-- explain
Expression ((Project names + (Projection + (Change column names to column identifiers + (Project names + Projection)))))
  Cube
    Aggregating
      Expression ((Before GROUP BY + (Change column names to column identifiers + (Project names + (Projection + DROP unused columns after JOIN)))))
        Join (JOIN FillRightFirst)
          Expression (Change column names to column identifiers)
            ReadFromStorage (SystemNumbers)
          Expression (Change column names to column identifiers)
            ReadFromStorage (SystemNumbers)
-- execute
0
2
1
0
-- GROUP BY WITH TOTALS before DISTINCT with on different columns => do _not_ remove DISTINCT
-- query
SELECT DISTINCT c
FROM
(
    SELECT
        a,
        sum(b) AS c
    FROM
    (
        SELECT
            x.number AS a,
            y.number AS b
        FROM numbers(3) AS x, numbers(3, 3) AS y
    )
    GROUP BY a WITH TOTALS
)
-- explain
Expression (Project names)
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression ((Projection + (Change column names to column identifiers + (Project names + Projection))))
        TotalsHaving
          Aggregating
            Expression ((Before GROUP BY + (Change column names to column identifiers + (Project names + (Projection + DROP unused columns after JOIN)))))
              Join (JOIN FillRightFirst)
                Expression (Change column names to column identifiers)
                  ReadFromStorage (SystemNumbers)
                Expression (Change column names to column identifiers)
                  ReadFromStorage (SystemNumbers)
-- execute
12

36
-- GROUP BY WITH TOTALS before DISTINCT with on the same columns => remove DISTINCT
-- query
SELECT DISTINCT a
FROM
(
    SELECT
        a,
        sum(b) AS c
    FROM
    (
        SELECT
            x.number AS a,
            y.number AS b
        FROM numbers(3) AS x, numbers(3, 3) AS y
    )
    GROUP BY a WITH TOTALS
)
-- explain
Expression ((Project names + (Projection + (Change column names to column identifiers + (Project names + Projection)))))
  TotalsHaving
    Aggregating
      Expression ((Before GROUP BY + (Change column names to column identifiers + (Project names + (Projection + DROP unused columns after JOIN)))))
        Join (JOIN FillRightFirst)
          Expression (Change column names to column identifiers)
            ReadFromStorage (SystemNumbers)
          Expression (Change column names to column identifiers)
            ReadFromStorage (SystemNumbers)
-- execute
0
2
1

0
-- DISTINCT COUNT() with GROUP BY => do _not_ remove DISTINCT
-- query
select distinct count() from numbers(10) group by number
-- explain
Expression (Project names)
  Distinct (DISTINCT)
    Distinct (Preliminary DISTINCT)
      Expression (Projection)
        Aggregating
          Expression ((Before GROUP BY + Change column names to column identifiers))
            ReadFromStorage (SystemNumbers)
-- execute
1