在 SQL 服务器中使用 Pivot 将列转换为行
Converting Columns Into Rows Using Pivot In SQL Server
我有如下table如下:
我想按如下方式转换:
+------------+---------------------+---------------------+
| Child_Code | SewingStart | SewingEnd |
+------------+---------------------+---------------------+
| 000001 | 2017-02-21 00:00:00 | 2017-03-21 00:00:00 |
+------------+---------------------+---------------------+
请帮忙!!
如果行数有限,可以使用条件聚合或数据透视表。但是,为此您需要一个专栏。所以:
select child_code,
max(case when seqnum = 1 then plan_date end) as plan_date_1,
max(case when seqnum = 2 then plan_date end) as plan_date_2
from (select t.*,
row_number() over (partition by child_code order by plan_date) as seqnum
from t
) t
group by child_code;
如果您知道所需的最大计划日期数,则只能使用此方法。否则,您将需要使用动态枢轴。思路是一样的,但是需要构造查询字符串,然后传递给sp_executesql
.
编辑:
如果您只有两个值,那么 group by
可能更容易。以下处理只有一个值的情况:
select child_code, min(plan_date) as plan_date_1,
(case when min(plan_date) <> max(plan_date) then max(plan_date)
end) as plan_date_2
from t
group by child_code;
如果 plan_date
的最大数量未知,您将需要使用动态 sql。您将需要使用 row_number()
to number each list partitioned by Child_Code
for use with pivot()
.
测试设置:
create table t (child_code varchar(6), plan_date datetime);
insert into t values ('000001','20170221'),('000001','20170321');
declare @cols nvarchar(max);
declare @sql nvarchar(max);
select @cols = stuff((
select distinct
',' + quotename('Plan_Date_'
+convert(nvarchar(10),row_number() over (
partition by Child_Code
order by Plan_Date
))
)
from t
for xml path (''), type).value('.','nvarchar(max)')
,1,1,'');
select @sql = '
select Child_Code, ' + @cols + '
from (
select
Child_Code
, Plan_Date
, rn=''Plan_Date_''+convert(nvarchar(10),row_number() over (
partition by Child_Code
order by Plan_Date
))
from t
) as a
pivot (max([Plan_Date]) for [rn] in (' + @cols + ') ) p';
select @sql as CodeGenerated;
exec sp_executesql @sql;
rextester 演示:http://rextester.com/YQCR87525
代码生成:
select Child_Code, [Plan_Date_1],[Plan_Date_2]
from (
select
Child_Code
, Plan_Date
, rn='Plan_Date_'+convert(nvarchar(10),row_number() over (
partition by Child_Code
order by Plan_Date
))
from t
) as a
pivot (max([Plan_Date]) for [rn] in ([Plan_Date_1],[Plan_Date_2]) ) p
returns
+------------+---------------------+---------------------+
| Child_Code | Plan_Date_1 | Plan_Date_2 |
+------------+---------------------+---------------------+
| 000001 | 21.02.2017 00:00:00 | 21.03.2017 00:00:00 |
+------------+---------------------+---------------------+
我有如下table如下:
我想按如下方式转换:
+------------+---------------------+---------------------+
| Child_Code | SewingStart | SewingEnd |
+------------+---------------------+---------------------+
| 000001 | 2017-02-21 00:00:00 | 2017-03-21 00:00:00 |
+------------+---------------------+---------------------+
请帮忙!!
如果行数有限,可以使用条件聚合或数据透视表。但是,为此您需要一个专栏。所以:
select child_code,
max(case when seqnum = 1 then plan_date end) as plan_date_1,
max(case when seqnum = 2 then plan_date end) as plan_date_2
from (select t.*,
row_number() over (partition by child_code order by plan_date) as seqnum
from t
) t
group by child_code;
如果您知道所需的最大计划日期数,则只能使用此方法。否则,您将需要使用动态枢轴。思路是一样的,但是需要构造查询字符串,然后传递给sp_executesql
.
编辑:
如果您只有两个值,那么 group by
可能更容易。以下处理只有一个值的情况:
select child_code, min(plan_date) as plan_date_1,
(case when min(plan_date) <> max(plan_date) then max(plan_date)
end) as plan_date_2
from t
group by child_code;
如果 plan_date
的最大数量未知,您将需要使用动态 sql。您将需要使用 row_number()
to number each list partitioned by Child_Code
for use with pivot()
.
测试设置:
create table t (child_code varchar(6), plan_date datetime);
insert into t values ('000001','20170221'),('000001','20170321');
declare @cols nvarchar(max);
declare @sql nvarchar(max);
select @cols = stuff((
select distinct
',' + quotename('Plan_Date_'
+convert(nvarchar(10),row_number() over (
partition by Child_Code
order by Plan_Date
))
)
from t
for xml path (''), type).value('.','nvarchar(max)')
,1,1,'');
select @sql = '
select Child_Code, ' + @cols + '
from (
select
Child_Code
, Plan_Date
, rn=''Plan_Date_''+convert(nvarchar(10),row_number() over (
partition by Child_Code
order by Plan_Date
))
from t
) as a
pivot (max([Plan_Date]) for [rn] in (' + @cols + ') ) p';
select @sql as CodeGenerated;
exec sp_executesql @sql;
rextester 演示:http://rextester.com/YQCR87525
代码生成:
select Child_Code, [Plan_Date_1],[Plan_Date_2]
from (
select
Child_Code
, Plan_Date
, rn='Plan_Date_'+convert(nvarchar(10),row_number() over (
partition by Child_Code
order by Plan_Date
))
from t
) as a
pivot (max([Plan_Date]) for [rn] in ([Plan_Date_1],[Plan_Date_2]) ) p
returns
+------------+---------------------+---------------------+
| Child_Code | Plan_Date_1 | Plan_Date_2 |
+------------+---------------------+---------------------+
| 000001 | 21.02.2017 00:00:00 | 21.03.2017 00:00:00 |
+------------+---------------------+---------------------+