-- Tags: long SET allow_experimental_analyzer = 1; -- Too slow SET max_bytes_before_external_sort = 0; SET max_bytes_before_external_group_by = 0; -- { echo } -- just something basic select number, count() over (partition by intDiv(number, 3) order by number rows unbounded preceding) from numbers(10); -- proper calculation across blocks select number, max(number) over (partition by intDiv(number, 3) order by number desc rows unbounded preceding) from numbers(10) settings max_block_size = 2; -- not a window function select number, abs(number) over (partition by toString(intDiv(number, 3)) rows unbounded preceding) from numbers(10); -- { serverError UNKNOWN_AGGREGATE_FUNCTION } -- no partition by select number, avg(number) over (order by number rows unbounded preceding) from numbers(10); -- no order by select number, quantileExact(number) over (partition by intDiv(number, 3) AS value order by number rows unbounded preceding) from numbers(10); -- can add an alias after window spec select number, quantileExact(number) over (partition by intDiv(number, 3) AS value order by number rows unbounded preceding) q from numbers(10); -- now we should be able to compute expressions with window functions select number, q * 10, quantileExact(number) over (partition by intDiv(number, 3) order by number rows unbounded preceding) q from numbers(10) order by number; -- must work in WHERE if you wrap it in a subquery select * from (select count(*) over (rows unbounded preceding) c from numbers(3)) where c > 0; -- should work in ORDER BY select number, max(number) over (partition by intDiv(number, 3) order by number desc rows unbounded preceding) m from numbers(10) order by m desc, number; -- also works in ORDER BY if you wrap it in a subquery select * from (select count(*) over (rows unbounded preceding) c from numbers(3)) order by c; -- Example with window function only in ORDER BY. Here we make a rank of all -- numbers sorted descending, and then sort by this rank descending, and must get -- the ascending order. select * from (select * from numbers(5) order by rand()) order by count() over (order by number desc rows unbounded preceding) desc; -- Aggregate functions as window function arguments. This query is semantically -- the same as the above one, only we replace `number` with -- `any(number) group by number` and so on. select * from (select * from numbers(5) order by rand()) group by number order by sum(any(number + 1)) over (order by min(number) desc rows unbounded preceding) desc; -- some more simple cases w/aggregate functions select sum(any(number)) over (rows unbounded preceding) from numbers(1); select sum(any(number) + 1) over (rows unbounded preceding) from numbers(1); select sum(any(number + 1)) over (rows unbounded preceding) from numbers(1); -- different windows -- an explain test would also be helpful, but it's too immature now and I don't -- want to change reference all the time select number, max(number) over (partition by intDiv(number, 3) order by number desc rows unbounded preceding), count(number) over (partition by intDiv(number, 5) order by number rows unbounded preceding) as m from numbers(31) order by number settings max_block_size = 2; -- two functions over the same window -- an explain test would also be helpful, but it's too immature now and I don't -- want to change reference all the time select number, max(number) over (partition by intDiv(number, 3) order by number desc rows unbounded preceding), count(number) over (partition by intDiv(number, 3) order by number desc rows unbounded preceding) as m from numbers(7) order by number settings max_block_size = 2; -- check that we can work with constant columns select median(x) over (partition by x) from (select 1 x); -- an empty window definition is valid as well select groupArray(number) over (rows unbounded preceding) from numbers(3); select groupArray(number) over () from numbers(3); -- This one tests we properly process the window function arguments. -- Seen errors like 'column `1` not found' from count(1). select count(1) over (rows unbounded preceding), max(number + 1) over () from numbers(3); -- Should work in DISTINCT select distinct sum(0) over (rows unbounded preceding) from numbers(2); select distinct any(number) over (rows unbounded preceding) from numbers(2); -- Various kinds of aliases are properly substituted into various parts of window -- function definition. with number + 1 as x select intDiv(number, 3) as y, sum(x + y) over (partition by y order by x rows unbounded preceding) from numbers(7); -- WINDOW clause select 1 window w1 as (); select sum(number) over w1, sum(number) over w2 from numbers(10) window w1 as (rows unbounded preceding), w2 as (partition by intDiv(number, 3) as value order by number rows unbounded preceding) ; -- FIXME both functions should use the same window, but they don't. Add an -- EXPLAIN test for this. select sum(number) over w1, sum(number) over (partition by intDiv(number, 3) as value order by number rows unbounded preceding) from numbers(10) window w1 as (partition by intDiv(number, 3) rows unbounded preceding) ; -- RANGE frame -- It's the default select sum(number) over () from numbers(3); -- Try some mutually prime sizes of partition, group and block, for the number -- of rows that is their least common multiple + 1, so that we see all the -- interesting corner cases. select number, intDiv(number, 3) p, mod(number, 2) o, count(number) over w as c from numbers(31) window w as (partition by p order by o, number range unbounded preceding) order by number settings max_block_size = 5 ; select number, intDiv(number, 5) p, mod(number, 3) o, count(number) over w as c from numbers(31) window w as (partition by p order by o, number range unbounded preceding) order by number settings max_block_size = 2 ; select number, intDiv(number, 5) p, mod(number, 2) o, count(number) over w as c from numbers(31) window w as (partition by p order by o, number range unbounded preceding) order by number settings max_block_size = 3 ; select number, intDiv(number, 3) p, mod(number, 5) o, count(number) over w as c from numbers(31) window w as (partition by p order by o, number range unbounded preceding) order by number settings max_block_size = 2 ; select number, intDiv(number, 2) p, mod(number, 5) o, count(number) over w as c from numbers(31) window w as (partition by p order by o, number range unbounded preceding) order by number settings max_block_size = 3 ; select number, intDiv(number, 2) p, mod(number, 3) o, count(number) over w as c from numbers(31) window w as (partition by p order by o range unbounded preceding) order by number settings max_block_size = 5 ; -- A case where the partition end is in the current block, and the frame end -- is triggered by the partition end. select min(number) over (partition by p) from (select number, intDiv(number, 3) p from numbers(10)); -- UNBOUNDED FOLLOWING frame end select min(number) over wa, min(number) over wo, max(number) over wa, max(number) over wo from (select number, intDiv(number, 3) p, mod(number, 5) o from numbers(31)) window wa as (partition by p order by o range between unbounded preceding and unbounded following), wo as (partition by p order by o rows between unbounded preceding and unbounded following) settings max_block_size = 2; -- ROWS offset frame start select number, p, count(*) over (partition by p order by number rows between 1 preceding and unbounded following), count(*) over (partition by p order by number rows between current row and unbounded following), count(*) over (partition by p order by number rows between 1 following and unbounded following) from (select number, intDiv(number, 5) p from numbers(31)) order by p, number settings max_block_size = 2; -- ROWS offset frame start and end select number, p, count(*) over (partition by p order by number rows between 2 preceding and 2 following) from (select number, intDiv(number, 7) p from numbers(71)) order by p, number settings max_block_size = 2; SELECT count(*) OVER (ROWS BETWEEN 2 PRECEDING AND CURRENT ROW) FROM numbers(4); -- frame boundaries that runs into the partition end select count() over (partition by intDiv(number, 3) rows between 100 following and unbounded following), count() over (partition by intDiv(number, 3) rows between current row and 100 following) from numbers(10); -- seen a use-after-free under MSan in this query once SELECT number, max(number) OVER (PARTITION BY intDiv(number, 7) ORDER BY number ASC NULLS LAST ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) FROM numbers(1024) SETTINGS max_block_size = 2 FORMAT Null; -- a corner case select count() over (); -- RANGE CURRENT ROW frame start select number, p, o, count(*) over (partition by p order by o range between current row and unbounded following) from (select number, intDiv(number, 5) p, mod(number, 3) o from numbers(31)) order by p, o, number settings max_block_size = 2; select count(*) over (rows between current row and current row), count(*) over (range between current row and current row) from numbers(3); -- RANGE OFFSET -- a basic RANGE OFFSET frame select x, min(x) over w, max(x) over w, count(x) over w from ( select toUInt8(number) x from numbers(11)) window w as (order by x asc range between 1 preceding and 2 following) order by x; -- overflow conditions select x, min(x) over w, max(x) over w, count(x) over w from ( select toUInt8(if(mod(number, 2), toInt64(255 - intDiv(number, 2)), toInt64(intDiv(number, 2)))) x from numbers(10) ) window w as (order by x range between 1 preceding and 2 following) order by x; select x, min(x) over w, max(x) over w, count(x) over w from ( select toInt8(multiIf( mod(number, 3) == 0, toInt64(intDiv(number, 3)), mod(number, 3) == 1, toInt64(127 - intDiv(number, 3)), toInt64(-128 + intDiv(number, 3)))) x from numbers(15) ) window w as (order by x range between 1 preceding and 2 following) order by x; -- We need large offsets to trigger overflow to positive direction, or -- else the frame end runs into partition end w/o overflow and doesn't move -- after that. The frame from this query is equivalent to the entire partition. select x, min(x) over w, max(x) over w, count(x) over w from ( select toUInt8(if(mod(number, 2), toInt64(255 - intDiv(number, 2)), toInt64(intDiv(number, 2)))) x from numbers(10) ) window w as (order by x range between 255 preceding and 255 following) order by x; -- RANGE OFFSET ORDER BY DESC select x, min(x) over w, max(x) over w, count(x) over w from ( select toUInt8(number) x from numbers(11)) t window w as (order by x desc range between 1 preceding and 2 following) order by x settings max_block_size = 1; select x, min(x) over w, max(x) over w, count(x) over w from ( select toUInt8(number) x from numbers(11)) t window w as (order by x desc range between 1 preceding and unbounded following) order by x settings max_block_size = 2; select x, min(x) over w, max(x) over w, count(x) over w from ( select toUInt8(number) x from numbers(11)) t window w as (order by x desc range between unbounded preceding and 2 following) order by x settings max_block_size = 3; select x, min(x) over w, max(x) over w, count(x) over w from ( select toUInt8(number) x from numbers(11)) t window w as (order by x desc range between unbounded preceding and 2 preceding) order by x settings max_block_size = 4; -- Check that we put windows in such an order that we can reuse the sort. -- First, check that at least the result is correct when we have many windows -- with different sort order. select number, count(*) over (partition by p order by number), count(*) over (partition by p order by number, o), count(*) over (), count(*) over (order by number), count(*) over (order by o), count(*) over (order by o, number), count(*) over (order by number, o), count(*) over (partition by p order by o, number), count(*) over (partition by p), count(*) over (partition by p order by o), count(*) over (partition by p, o order by number) from (select number, intDiv(number, 3) p, mod(number, 5) o from numbers(16)) t order by number ; -- The EXPLAIN for the above query would be difficult to understand, so check some -- simple cases instead. explain select count(*) over (partition by p), count(*) over (), count(*) over (partition by p order by o) from (select number, intDiv(number, 3) p, mod(number, 5) o from numbers(16)) t ; explain select count(*) over (order by o, number), count(*) over (order by number) from (select number, intDiv(number, 3) p, mod(number, 5) o from numbers(16)) t ; -- A test case for the sort comparator found by fuzzer. SELECT max(number) OVER (ORDER BY number DESC NULLS FIRST), max(number) OVER (ORDER BY number ASC NULLS FIRST) FROM numbers(2) ; -- optimize_read_in_order conflicts with sorting for window functions, check that -- it is disabled. drop table if exists window_mt; create table window_mt engine MergeTree order by number as select number, mod(number, 3) p from numbers(100); select number, count(*) over (partition by p) from window_mt order by number limit 10 settings optimize_read_in_order = 0; select number, count(*) over (partition by p) from window_mt order by number limit 10 settings optimize_read_in_order = 1; drop table window_mt; -- some true window functions -- rank and friends select number, p, o, count(*) over w, rank() over w, dense_rank() over w, row_number() over w from (select number, intDiv(number, 5) p, mod(number, 3) o from numbers(31) order by o, number) t window w as (partition by p order by o, number) order by p, o, number settings max_block_size = 2; -- our replacement for lag/lead select anyOrNull(number) over (order by number rows between 1 preceding and 1 preceding), anyOrNull(number) over (order by number rows between 1 following and 1 following) from numbers(5); -- variants of lag/lead that respect the frame select number, p, pp, lagInFrame(number) over w as lag1, lagInFrame(number, number - pp) over w as lag2, lagInFrame(number, number - pp, number * 11) over w as lag, leadInFrame(number, number - pp, number * 11) over w as lead from (select number, intDiv(number, 5) p, p * 5 pp from numbers(16)) window w as (partition by p order by number rows between unbounded preceding and unbounded following) order by number settings max_block_size = 3; ; -- careful with auto-application of Null combinator select lagInFrame(toNullable(1)) over (); select lagInFrameOrNull(1) over (); -- { serverError BAD_ARGUMENTS } -- this is the same as `select max(Null::Nullable(Nothing))` select intDiv(1, NULL) x, toTypeName(x), max(x) over (); -- to make lagInFrame return null for out-of-frame rows, cast the argument to -- Nullable; otherwise, it returns default values. SELECT number, lagInFrame(toNullable(number), 1) OVER w, lagInFrame(toNullable(number), 2) OVER w, lagInFrame(number, 1) OVER w, lagInFrame(number, 2) OVER w FROM numbers(4) WINDOW w AS (ORDER BY number ASC) ; -- case-insensitive SQL-standard synonyms for any and anyLast select number, fIrSt_VaLue(number) over w, lAsT_vAlUe(number) over w from numbers(10) window w as (order by number range between 1 preceding and 1 following) order by number ; -- nth_value without specific frame range given select number, nth_value(number, 1) over w as firstValue, nth_value(number, 2) over w as secondValue, nth_value(number, 3) over w as thirdValue, nth_value(number, 4) over w as fourthValue from numbers(10) window w as (order by number) order by number ; -- nth_value with frame range specified select number, nth_value(number, 1) over w as firstValue, nth_value(number, 2) over w as secondValue, nth_value(number, 3) over w as thirdValue, nth_value(number, 4) over w as fourthValue from numbers(10) window w as (order by number range between 1 preceding and 1 following) order by number ; -- to make nth_value return null for out-of-frame rows, cast the argument to -- Nullable; otherwise, it returns default values. SELECT number, nth_value(toNullable(number), 1) OVER w as firstValue, nth_value(toNullable(number), 3) OVER w as thridValue FROM numbers(5) WINDOW w AS (ORDER BY number ASC) ; -- nth_value UBsan SELECT nth_value(1, -1) OVER (); -- { serverError BAD_ARGUMENTS } SELECT nth_value(1, 0) OVER (); -- { serverError BAD_ARGUMENTS } SELECT nth_value(1, /* INT64_MAX+1 */ 0x7fffffffffffffff+1) OVER (); -- { serverError BAD_ARGUMENTS } SELECT nth_value(1, /* INT64_MAX */ 0x7fffffffffffffff) OVER (); SELECT nth_value(1, 1) OVER (); -- lagInFrame UBsan SELECT lagInFrame(1, -1) OVER (); -- { serverError BAD_ARGUMENTS } SELECT lagInFrame(1, 0) OVER (); SELECT lagInFrame(1, /* INT64_MAX+1 */ 0x7fffffffffffffff+1) OVER (); -- { serverError BAD_ARGUMENTS } SELECT lagInFrame(1, /* INT64_MAX */ 0x7fffffffffffffff) OVER (); SELECT lagInFrame(1, 1) OVER (); -- leadInFrame UBsan SELECT leadInFrame(1, -1) OVER (); -- { serverError BAD_ARGUMENTS } SELECT leadInFrame(1, 0) OVER (); SELECT leadInFrame(1, /* INT64_MAX+1 */ 0x7fffffffffffffff+1) OVER (); -- { serverError BAD_ARGUMENTS } SELECT leadInFrame(1, /* INT64_MAX */ 0x7fffffffffffffff) OVER (); SELECT leadInFrame(1, 1) OVER (); -- nth_value Msan SELECT nth_value(1, '') OVER (); -- { serverError BAD_ARGUMENTS } -- lagInFrame Msan SELECT lagInFrame(1, '') OVER (); -- { serverError BAD_ARGUMENTS } -- leadInFrame Msan SELECT leadInFrame(1, '') OVER (); -- { serverError BAD_ARGUMENTS } -- In this case, we had a problem with PartialSortingTransform returning zero-row -- chunks for input chunks w/o columns. select count() over () from numbers(4) where number < 2; -- floating point RANGE frame select count(*) over (order by toFloat32(number) range 5 preceding), count(*) over (order by toFloat64(number) range 5 preceding), count(*) over (order by toFloat32(number) range between current row and 5 following), count(*) over (order by toFloat64(number) range between current row and 5 following) from numbers(7) ; -- negative offsets should not be allowed select count() over (order by toInt64(number) range between -1 preceding and unbounded following) from numbers(1); -- { serverError BAD_ARGUMENTS } select count() over (order by toInt64(number) range between -1 following and unbounded following) from numbers(1); -- { serverError BAD_ARGUMENTS } select count() over (order by toInt64(number) range between unbounded preceding and -1 preceding) from numbers(1); -- { serverError BAD_ARGUMENTS } select count() over (order by toInt64(number) range between unbounded preceding and -1 following) from numbers(1); -- { serverError BAD_ARGUMENTS } -- a test with aggregate function that allocates memory in arena select sum(a[length(a)]) from ( select groupArray(number) over (partition by modulo(number, 11) order by modulo(number, 1111), number) a from numbers_mt(10000) ) settings max_block_size = 7; -- a test with aggregate function which is -state type select bitmapCardinality(bs) from ( select groupBitmapMergeState(bm) over (order by k asc rows between unbounded preceding and current row) as bs from ( select groupBitmapState(number) as bm, k from ( select number, number % 3 as k from numbers(3) ) group by k ) ); -- -INT_MIN row offset that can lead to problems with negation, found when fuzzing -- under UBSan. Should be limited to at most INT_MAX. select count() over (rows between 2147483648 preceding and 2147493648 following) from numbers(2); -- { serverError BAD_ARGUMENTS } -- Somehow in this case WindowTransform gets empty input chunks not marked as -- input end, and then two (!) empty input chunks marked as input end. Whatever. select count() over () from (select 1 a) l inner join (select 2 a) r using a; -- This case works as expected, one empty input chunk marked as input end. select count() over () where null; -- Inheriting another window. select number, count() over (w1 rows unbounded preceding) from numbers(10) window w0 as (partition by intDiv(number, 5) as p), w1 as (w0 order by mod(number, 3) as o, number) order by p, o, number ; -- can't redefine PARTITION BY select count() over (w partition by number) from numbers(1) window w as (partition by intDiv(number, 5)); -- { serverError BAD_ARGUMENTS } -- can't redefine existing ORDER BY select count() over (w order by number) from numbers(1) window w as (partition by intDiv(number, 5) order by mod(number, 3)); -- { serverError BAD_ARGUMENTS } -- parent window can't have frame select count() over (w range unbounded preceding) from numbers(1) window w as (partition by intDiv(number, 5) order by mod(number, 3) rows unbounded preceding); -- { serverError BAD_ARGUMENTS } -- looks weird but probably should work -- this is a window that inherits and changes nothing select count() over (w) from numbers(1) window w as (); -- nonexistent parent window select count() over (w2 rows unbounded preceding); -- { serverError BAD_ARGUMENTS }