select 每个 window 分区上的不同名称和用户名

Question

我在 presto 上使用 window 函数来获取每个组的不同名称和用户名行。我确实在我的 name 列上应用了 ROW_NUMBER() 作为分区的 datetime 列作为排序依据，但我得到的结果低于结果

当前输出：

name    top_user     Count           Date         Price  Percent  Volume      username   
ENZC    1            5   2021-03-07 08:11:14.000 0.1189  45.05   86131409    DeviantImmortal
ENZC    5            5   2021-03-07 08:11:14.000 0.1189  45.05   86131409    OtcRock
ENZC    2            5   2021-03-07 08:11:14.000 0.1189  45.05   86131409    OtcRock
ENZC    3            5   2021-03-07 08:11:14.000 0.1189  45.05   86131409    STOCKAHOLIC55
ENZC    4            5   2021-03-07 08:11:14.000 0.1189  45.05   86131409    stockpro20
HCMC    3            5   2021-03-07 08:34:33.000 0.0002  15.2    1376689232  Barta57
HCMC    5            5   2021-03-07 08:34:33.000 0.0002  15.2    1376689232  PennyProfitPro
HCMC    2            5   2021-03-07 08:34:33.000 0.0002  15.2    1376689232  Stocktipstoday1
HCMC    1            5   2021-03-07 08:34:33.000 0.0002  15.2    1376689232  TTrader1976
HCMC    4            5   2021-03-07 08:34:33.000 0.0002  15.2    1376689232  stockpro20
HQGE    5            6   2021-03-07 07:40:38.000 -0.0017 -16.04  63596752    BerkshireCapGrp
HQGE    1            6   2021-03-07 07:40:38.000 -0.0017 -16.04  63596752    OwnThePlayOTC
HQGE    2            6   2021-03-07 07:40:38.000 -0.0017 -16.04  63596752    PennyStockGeeks
HQGE    3            6   2021-03-07 07:40:38.000 -0.0017 -16.04  63596752    TaylorB16445829
HQGE    4            6   2021-03-07 07:40:38.000 -0.0017 -16.04  63596752    iammpremm
LTNC    2            8   2021-03-07 08:33:19.000 0.0028  10.73   293126083   BigTawno
LTNC    5            8   2021-03-07 08:33:19.000 0.0028  10.73   293126083   Faith03777244
LTNC    3            8   2021-03-07 08:33:19.000 0.0028  10.73   293126083   OneTickMoline
LTNC    1            8   2021-03-07 08:33:19.000 0.0028  10.73   293126083   OneTickMoline
LTNC    4            8   2021-03-07 08:33:19.000 0.0028  10.73   293126083   Stock_Pop
OZSC    2           10  2021-03-07 08:34:38.000 0.0685  72.87   330616866   JZavitka
OZSC    3           10  2021-03-07 08:34:38.000 0.0685  72.87   330616866   JZavitka
OZSC    1           10  2021-03-07 08:34:38.000 0.0685  72.87   330616866   S_AnglinIV
OZSC    4           10  2021-03-07 08:34:38.000 0.0685  72.87   330616866   S_AnglinIV
OZSC    5           10  2021-03-07 08:34:38.000 0.0685  72.87   330616866   claydeath1
SANP    2           5   2021-03-07 08:11:38.000 0.0049  101.04  907907634   1deadmanx
SANP    3           5   2021-03-07 08:11:38.000 0.0049  101.04  907907634   BillTsamis
SANP    5           5   2021-03-07 08:11:38.000 0.0049  101.04  907907634   Fluffypillows9
SANP    1           5   2021-03-07 08:11:38.000 0.0049  101.04  907907634   Fluffypillows9

当前查询：

SELECT * FROM 
(
SELECT  name, username , datetime, message, 
ROW_NUMBER() OVER (PARTITION BY name ORDER BY datetime ASC) AS top_user FROM table_name 
)
WHERE top_user < 6 ORDER BY name

预期输出：为每个window分区获取不同的名称和用户名，即没有两行任何window 具有相同的名称和用户名

name top_user Count Date Price Percent Volume username ENZC 1 5 2021-03-07 08:11:14.000 0.1189 45.05 86131409 DeviantImmortal ENZC 2 5 2021-03-07 08:11:14.000 0.1189 45.05 86131409 OtcRock ENZC 3 5 2021-03-07 08:11:14.000 0.1189 45.05 86131409 STOCKAHOLIC55 ENZC 4 5 2021-03-07 08:11:14.000 0.1189 45.05 86131409 stockpro20 HCMC 3 5 2021-03-07 08:34:33.000 0.0002 15.2 1376689232 Barta57 HCMC 5 5 2021-03-07 08:34:33.000 0.0002 15.2 1376689232 PennyProfitPro HCMC 2 5 2021-03-07 08:34:33.000 0.0002 15.2 1376689232 Stocktipstoday1 HCMC 1 5 2021-03-07 08:34:33.000 0.0002 15.2 1376689232 TTrader1976 HCMC 4 5 2021-03-07 08:34:33.000 0.0002 15.2 1376689232 stockpro20 HQGE 5 6 2021-03-07 07:40:38.000 -0.0017 -16.04 63596752 BerkshireCapGrp HQGE 1 6 2021-03-07 07:40:38.000 -0.0017 -16.04 63596752 OwnThePlayOTC HQGE 2 6 2021-03-07 07:40:38.000 -0.0017 -16.04 63596752 PennyStockGeeks HQGE 3 6 2021-03-07 07:40:38.000 -0.0017 -16.04 63596752 TaylorB16445829 HQGE 4 6 2021-03-07 07:40:38.000 -0.0017 -16.04 63596752 iammpremm LTNC 2 8 2021-03-07 08:33:19.000 0.0028 10.73 293126083 BigTawno LTNC 5 8 2021-03-07 08:33:19.000 0.0028 10.73 293126083 Faith03777244 LTNC 3 8 2021-03-07 08:33:19.000 0.0028 10.73 293126083 OneTickMoline LTNC 4 8 2021-03-07 08:33:19.000 0.0028 10.73 293126083 Stock_Pop OZSC 2 10 2021-03-07 08:34:38.000 0.0685 72.87 330616866 JZavitka OZSC 1 10 2021-03-07 08:34:38.000 0.0685 72.87 330616866 S_AnglinIV OZSC 5 10 2021-03-07 08:34:38.000 0.0685 72.87 330616866 claydeath1 SANP 2 5 2021-03-07 08:11:38.000 0.0049 101.04 907907634 1deadmanx SANP 3 5 2021-03-07 08:11:38.000 0.0049 101.04 907907634 BillTsamis SANP 1 5 2021-03-07 08:11:38.000 0.0049 101.04 907907634 Fluffypillows9

Answer 1

一种方法是过滤掉重复的用户名，首先按名称和用户名进行分区，然后再次按名称进行分区：

SELECT *
FROM (
  SELECT *, ROW_NUMBER() OVER (PARTITION BY name ORDER BY datetime) rn2
  FROM (
    SELECT *, ROW_NUMBER() OVER (PARTITION BY name, username ORDER BY datetime) rn1
    FROM tablename
  ) t1
  WHERE t1.rn1 = 1 
) t2
WHERE t2.rn2 < 6

select 每个 window 分区上的不同名称和用户名

select distinct name and username on each window partition on presto

sql

distinct

window-functions

presto

amazon-athena