Flatten variable length DB rows into a single column - sql

I've searched and can't find a solution to this that exactly fits my needs, nor can I find one that I can modify. I have a database table, for simplicity we'll say it has three columns (packageID, carrier, and sequence). For any package there can be one or more carriers that have handled the package. I can do a query like
SELECT packageID, carrier
FROM packageFlow
ORDER BY sequence
to get a list of all the people that have handled packages that looks like:
packageID, carrier
1, Bob
1, Jim
1, Sally
1, Ron
2, Reggie
2, Mary
2, Bruce
What I need though is to get the results into rows that look like:
packageID|carrier1|carrier2|carrier3|carrier4
1 |Bob |Jim |Sally |Ron
2 |Reggie |Mary |Bruce
Pivot doesn't seem to do what I need since I'm not aggregating anything and I can't get a CTE to work correctly either. I'd appreciate any nudges in the right direction.

This data transformation is a PIVOT. Starting in SQL Server 2005, there is a function that will convert the rows into columns.
If you have a known number of values, then you can hard-code your query:
select *
from
(
select packageid, carrier,
'Carrier_'+cast(row_number() over(partition by packageid order by packageid) as varchar(10)) col
from packageflow
) src
pivot
(
max(carrier)
for col in (Carrier_1, Carrier_2, Carrier_3, Carrier_4)
) piv
See SQL Fiddle with Demo.
If you have an unknown number of Carrier values that you want to turn into columns, then you can use dynamic sql:
DECLARE #cols AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX)
select #cols = STUFF((SELECT distinct ',' + QUOTENAME(t.col)
from
(
select 'Carrier_'+cast(row_number() over(partition by packageid order by packageid) as varchar(10)) col
from packageFlow
) t
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query = 'SELECT packageid,' + #cols + ' from
(
select packageid, carrier,
''Carrier_''+cast(row_number() over(partition by packageid order by packageid) as varchar(10)) col
from packageflow
) x
pivot
(
max(carrier)
for col in (' + #cols + ')
) p '
execute(#query)
See SQL Fiddle with Demo.
Note: you will replace the order by packageid with order by sequence
The result of both queries is:
| PACKAGEID | CARRIER_1 | CARRIER_2 | CARRIER_3 | CARRIER_4 |
-------------------------------------------------------------
| 1 | Bob | Jim | Sally | Ron |
| 2 | Reggie | Mary | Bruce | (null) |

Related

Count number of values across multiple columns

I have a table with 11 columns. The first column includes the category names. The remaining 10 columns have values like white, green, big, damaged etc. and these values can change in time.
I need a SQL query to find how many are there in table (in 10 columns) each value.
Table 1:
+------------+------------+
| ID | decription |
+------------+------------+
| 1 | white |
| 2 | green |
| 3 | big |
| 4 | damaged |
+------------+------------+
Table 2:
+------------+-----------+-----------+-----------+
| CATEGORY | SECTION 1 | SECTION 2 | SECTION 3 |
+------------+-----------+-----------+-----------+
| Category 1 | white | green | big |
| Category 2 | big | damaged | white |
| Category 1 | white | green | big |
| Category 3 | big | damaged | white |
+------------+-----------+-----------+-----------+
Desired result:
+------------+-------+-------+-----+---------+
| CATEGORY | White | Green | Big | Damaged |
+------------+-------+-------+-----+---------+
| Category 1 | 20 | 10 | 9 | 50 |
| Category 2 | 25 | 21 | 15 | 5 |
+------------+-------+-------+-----+---------+
Is it possible doing like this dynamically just as query ?
its on MS sql in visual studio reporting
Thanks
You've got yourself a bit of a mess with the design and the desired result. The problem is that your table is denormalized and then the final result you want is also denormalized. You can get the final result by unpivoting your Section columns, then pivoting the values of those columns. You further add to the mess by needing to do this dynamically.
First, I'd advise you to rethink your table structure because this is far too messy to maintain.
In the meantime, before you even think about writing a dynamic version to get the result you have to get the logic correct via a static or hard-coded query. Now, you didn't state which version of SQL Server you are using but you first need to unpivot the Section columns. You can use either the UNPIVOT function or CROSS APPLY. Your query will start with something similar to the following:
select
category,
value
from yourtable
unpivot
(
value for cols in (Section1,Section2,Section3)
) u
See SQL Fiddle with Demo. This gets your data into the format:
| CATEGORY | VALUE |
|------------|---------|
| Category 1 | white |
| Category 1 | green |
| Category 1 | big |
| Category 2 | big |
| Category 2 | damaged |
| Category 2 | white |
Now you have multiple Category rows - one for each value that previously were in the Section columns. Since you want a total count of each word in the Category, you can now apply the pivot function:
select
category,
white, green, big, damaged
from
(
select
category,
value
from yourtable
unpivot
(
value for cols in (Section1,Section2,Section3)
) u
) un
pivot
(
count(value)
for value in (white, green, big, damaged)
) p;
See SQL Fiddle with Demo. This will give you the result that you want but now you need this to be done dynamically. You'll have to use dynamic SQL which will create a SQL string that will be executed giving you the final result.
If the number of columns to UNPIVOT is limited, then you will create a list of the new column values in a string and then execute it similar to:
DECLARE #query AS NVARCHAR(MAX),
#colsPivot as NVARCHAR(MAX);
select #colsPivot
= STUFF((SELECT ',' + quotename(SectionValue)
from yourtable
cross apply
(
select Section1 union all
select Section2 union all
select Section3
) d (SectionValue)
group by SectionValue
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query
= 'select category, '+#colspivot+'
from
(
select
category,
value
from yourtable
unpivot
(
value
for cols in (Section1, Section2, Section3)
) un
) x
pivot
(
count(value)
for value in ('+ #colspivot +')
) p'
exec sp_executesql #query
See SQL Fiddle with Demo
If you have an unknown number of columns to unpivot, then your process will be a bit more complicated. You'll need to generate a string with the columns to unpivot, you can use the sys.columns table to get this list:
select #colsUnpivot
= stuff((select ','+quotename(C.name)
from sys.columns as C
where C.object_id = object_id('yourtable') and
C.name like 'Section%'
for xml path('')), 1, 1, '')
Then you'll need to get a list of the new column values - but since these are dynamic we will need to generate this list with a bit of work. You'll need to unpivot the table to generate the list of values into a temporary table for use. Create a temp table to store the values:
create table #Category_Section
(
Category varchar(50),
SectionValue varchar(50)
);
Load the temp table with the data that you need to unpivot:
set #unpivotquery
= 'select
category,
value
from yourtable
unpivot
(
value for cols in ('+ #colsUnpivot +')
) u'
insert into #Category_Section exec(#unpivotquery);
See SQL Fiddle with Demo. You'll see that your data looks the same as the static version above. Now you need to create a string with the values from the temp table that will be used in the final query:
select #colsPivot
= STUFF((SELECT ',' + quotename(SectionValue)
from #Category_Section
group by SectionValue
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
Once you have all this you can put it together into a final query:
DECLARE #colsUnpivot AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX),
#colsPivot as NVARCHAR(MAX),
#unpivotquery AS NVARCHAR(MAX);
select #colsUnpivot
= stuff((select ','+quotename(C.name)
from sys.columns as C
where C.object_id = object_id('yourtable') and
C.name like 'Section%'
for xml path('')), 1, 1, '');
create table #Category_Section
(
Category varchar(50),
SectionValue varchar(50)
);
set #unpivotquery
= 'select
category,
value
from yourtable
unpivot
(
value for cols in ('+ #colsUnpivot +')
) u';
insert into #Category_Section exec(#unpivotquery);
select #colsPivot
= STUFF((SELECT ',' + quotename(SectionValue)
from #Category_Section
group by SectionValue
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query
= 'select category, '+#colspivot+'
from
(
select
category,
value
from yourtable
unpivot
(
value
for cols in ('+ #colsunpivot +')
) un
) x
pivot
(
count(value)
for value in ('+ #colspivot +')
) p'
exec sp_executesql #query
See SQL Fiddle with Demo. All versions will get you the end result:
| CATEGORY | BIG | DAMAGED | GREEN | WHITE |
|------------|-----|---------|-------|-------|
| Category 1 | 2 | 0 | 2 | 2 |
| Category 2 | 1 | 1 | 0 | 1 |
| Category 3 | 1 | 1 | 0 | 1 |
If your values are stored in a separate table, then you would generate your list of values from that table:
DECLARE #query AS NVARCHAR(MAX),
#colsPivot as NVARCHAR(MAX);
select #colsPivot
= STUFF((SELECT ',' + quotename(decription)
from descriptions
group by decription
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query
= 'select category, '+#colspivot+'
from
(
select
category,
value
from yourtable
unpivot
(
value
for cols in (Section1, Section2, Section3)
) un
) x
pivot
(
count(value)
for value in ('+ #colspivot +')
) p'
exec sp_executesql #query
See SQL Fiddle with Demo and still get the same result:
| CATEGORY | BIG | DAMAGED | GREEN | WHITE |
|------------|-----|---------|-------|-------|
| Category 1 | 2 | 0 | 2 | 2 |
| Category 2 | 1 | 1 | 0 | 1 |
| Category 3 | 1 | 1 | 0 | 1 |
select category,
SUM(CASE when section1='white' then 1 when section2='white' then 1 when section3='white' then 1 else 0 end) as white,
SUM(CASE when section1='green' then 1 when section2='green' then 1 when section3='green' then 1 else 0 end) as green,
SUM(CASE when section1='damaged' then 1 when section2='damaged' then 1 when section3='damaged' then 1 else 0 end) as damaged,
SUM(CASE when section1='big' then 1 when section2='big' then 1 when section3='big' then 1 else 0 end) as big
from test
group by category
SQLFiddle
You can extend more to n section values as shown above gor section1,section2,section3

Pivot Table issue: no identifier in original table or enumeration

I'm having an issue with using pivot or dynamic tables.
I found this example to be very helpful in understanding how I can accomplish this task; however, I am missing the inside identifier column.
Dynamic Pivot Columns in SQL Server
There is an SQL fiddle provided here in the post above: http://www.sqlfiddle.com/#!3/7fad2/6
You can see that in the second table propertyObjects, there is a count 1, 2, 3, 4 for each objectID. I do not have that propertyID count. This is all I have
case category
1 xx
1 xyx
1 abc
2 ghj
2 asdf
3 dfgh
As you can see I have a number of different categories for each case, but no category identifier field.
This is what I need:
case cat1 cat2 cat3
1 xx xyx abc
2 ghj asdf
3 dfgh
So I am thinking I might need to add a column to the source table and somehow enumerate the categories per case. This would make it possible for me to use the pivot in the provided example. Thoughts?
I tried to use row_number to accomplish this, but it does not stop at each case number, it just continues on counting the entire table.
Since you have multiple values for each case, then you will need to use row_number() to get the separate columns for each category.
Before you write the dynamic SQL version I would first write a hard-coded version. The code will be similar to:
SELECT [case], cat1, cat2, cat3
FROM
(
SELECT [case], category,
'cat'+
cast(row_number() over(partition by [case]
order by category) as varchar(10)) seq
FROM yourTable
) x
PIVOT
(
max(category)
for seq in (cat1, cat2, cat3)
)p;
See SQL Fiddle with Demo.
Now you have the logic down, then you can convert it to dynamic SQL:
DECLARE #cols AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX)
select #cols = STUFF((SELECT distinct ',' + QUOTENAME('cat'+cast(seq as varchar(10)))
from
(
select row_number() over(partition by [case]
order by category) seq
from yourtable
) d
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query = 'SELECT [case],' + #cols + '
from
(
SELECT [case], category,
''cat''+
cast(row_number() over(partition by [case]
order by category) as varchar(10)) seq
FROM yourTable
) x
pivot
(
max(category)
for seq in (' + #cols + ')
) p '
execute sp_executesql #query;
See SQL Fiddle with Demo. This will give you the result:
| CASE | CAT1 | CAT2 | CAT3 |
|------|------|--------|--------|
| 1 | abc | xx | xyx |
| 2 | asdf | ghj | (null) |
| 3 | dfgh | (null) | (null) |

PIVOTing variable number of rows to columns

I'm currently attempting to PIVOT some rows to columns. The problem is that I don't always know how many rows will be available. Let's look at an example:
Values_Table Columns_Table
------------ -----------
ID ID
ColumnsTableID GroupID
Value ColumnName
RESULTS"
Columns_Table
---------------
ID | GroupID | ColumnName
---------------------------------
0 1 Cats
1 1 Dogs
2 1 Birds
3 2 Pontiac
4 2 Ford
5 3 Trex
6 3 Raptor
7 3 Triceratops
8 3 Kentrosaurus
SQL FIDDLE EXAMPLE of a STATIC pivot. I am trying to achieve a dynamic pivot - http://sqlfiddle.com/#!3/2be82/1
So, here is my dilemma: I want to be able to pivot an unknown number of columns based on, in this scenario, the GroupID.
I want to be able to PIVOT, for example, all the rows in GroupID 3 into columns. I would need to do this without knowing how many rows are in groupID 3.
The design of the database is set in stone, so I can't do anything about that. All I can do is work with what I have :(
So, that said- does anyone have any suggestions on how to accomplish this task of PIVOTing an unknown number of rows into columns based on, in this example, the groupID?
If you are not going to know the values ahead of time, then you will need to look at using dynamic SQL. This will create a SQL String that will be executed, this is required because the list of columns must be known when the query is run.
The code will be similar to:
DECLARE #cols AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX),
#groupid as int
set #groupid = 3
select #cols = STUFF((SELECT distinct ',' + QUOTENAME(GroupName)
from Columns_Table
where groupid = #groupid
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query = 'SELECT ' + #cols + '
from
(
SELECT B.GroupName, A.Value
, row_number() over(partition by a.ColumnsTableID
order by a.Value) seq
FROM Values_Table AS A
INNER JOIN Columns_Table AS B
ON A.ColumnsTableID = B.ID
where b.groupid = '+cast(#groupid as varchar(10))+'
) p
pivot
(
min(P.Value)
for P.GroupName in (' + #cols + ')
) p '
execute sp_executesql #query;
See SQL Fiddle with Demo. For groupid of 3, then the result will be:
| KENTROSAURUS | RAPTOR | TREX | TRICERATOPS |
| whatisthiseven | Itsaraptor | Jurassic | landbeforetime |
| (null) | zomg | Park | (null) |

How do I Pivot Vertical Data to Horizontal Data SQL with Variable Row Lengths?

Okay I have the following table.
Name ID Website
Aaron | 2305 | CoolSave1
Aaron | 8464 | DiscoWorld1
Adriana | 2956 | NewCin1
Adriana | 5991 | NewCin2
Adriana | 4563 NewCin3
I would like to transform it into the following way.
Adriana | 2956 | NewCin1 | 5991 | NewCin2 | 4563 | NewCin3
Aaron | 2305 | CoolSave1 | 8464 | DiscoWorld | NULL | NULL
As you can see i am trying to take the first name from the first table and make a single row with all the IDs / Websites associated with that name. The problem is, there is a variable amount of websites that may be associated with each name. To handle this i'd like to just make a table with with the number of fields sequal to the max line item, and then for the subsequent lineitems, plug in a NULL where there are not enough data.
In order to get the result, you will need to apply both the UNPIVOT and the PIVOT functions to the data. The UNPIVOT will take the columns (ID, website) and convert them to rows, once this is done, then you can PIVOT the data back into columns.
The UNPIVOT code will be similar to the following:
select name,
col+'_'+cast(col_num as varchar(10)) col,
value
from
(
select name,
cast(id as varchar(11)) id,
website,
row_number() over(partition by name order by id) col_num
from yt
) src
unpivot
(
value
for col in (id, website)
) unpiv;
See SQL Fiddle with Demo. This gives a result:
| NAME | COL | VALUE |
-------------------------------------
| Aaron | id_1 | 2305 |
| Aaron | website_1 | CoolSave1 |
| Aaron | id_2 | 8464 |
| Aaron | website_2 | DiscoWorld1 |
As you can see I applied a row_number() to the data prior to the unpivot, the row number is used to generate the new column names. The columns in the UNPIVOT must also be of the same datatype, I applied a cast to the id column in the subquery to convert the data to a varchar prior to the pivot.
The col values are then used in the PIVOT. Once the data has been unpivoted, you apply the PIVOT function:
select *
from
(
select name,
col+'_'+cast(col_num as varchar(10)) col,
value
from
(
select name,
cast(id as varchar(11)) id,
website,
row_number() over(partition by name order by id) col_num
from yt
) src
unpivot
(
value
for col in (id, website)
) unpiv
) d
pivot
(
max(value)
for col in (id_1, website_1, id_2, website_2, id_3, website_3)
) piv;
See SQL Fiddle with Demo.
The above version works great if you have a limited or known number of values. But if the number of rows is unknown, then you will need to use dynamic SQL to generate the result:
DECLARE #cols AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX)
select #cols = STUFF((SELECT ',' + QUOTENAME( col+'_'+cast(col_num as varchar(10)))
from
(
select row_number() over(partition by name order by id) col_num
from yt
) t
cross apply
(
select 'id' col union all
select 'website'
) c
group by col, col_num
order by col_num, col
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query = 'SELECT name,' + #cols + '
from
(
select name,
col+''_''+cast(col_num as varchar(10)) col,
value
from
(
select name,
cast(id as varchar(11)) id,
website,
row_number() over(partition by name order by id) col_num
from yt
) src
unpivot
(
value
for col in (id, website)
) unpiv
) x
pivot
(
max(value)
for col in (' + #cols + ')
) p '
execute(#query);
See SQL Fiddle with Demo. Both versions give the result:
| NAME | ID_1 | WEBSITE_1 | ID_2 | WEBSITE_2 | ID_3 | WEBSITE_3 |
------------------------------------------------------------------------
| Aaron | 2305 | CoolSave1 | 8464 | DiscoWorld1 | (null) | (null) |
| Adriana | 2956 | NewCin1 | 4563 | NewCin3 | 5991 | NewCin2 |

SQL Server - Rows to Columns without Aggregation

I have data that looks like this:
address | id
12AnyStreet | 1234
12AnyStreet | 1235
12AnyStreet | 1236
12AnyStreet | 1237
My goal is to make it look like this:
Address id1 id2 id3 id4
123Any 1234 1235 1246 1237
Based on some Googling and what not, I was able to generate the following CTE:
with cust_cte (Address, id, RID) as (
SELECT Address, id,
ROW_NUMBER() OVER (PARTITION BY (Address) ORDER BY Address) AS RID
FROM tab)
The next step would be to pivot so that for each RID, I place the associated id in the column. However, I can't seem to get the example I found to work. Rather than post the rest of the example which may not even really apply, I'll leave it up to the audience. Other novel approaches not necessarily utilizing the CTE are also appreciated. This will be chugging through a lot of data, so efficiency is important.
You can transform this data using the PIVOT function in SQL Server. In order to PIVOT the data, you will want to create your new column using the row_number().
If you have a known number of values, then you can hard-code the query:
select *
from
(
select address, id,
'id_'+cast(row_number() over(partition by address
order by id) as varchar(20)) rn
from yourtable
) src
pivot
(
max(id)
for rn in ([id_1], [id_2], [id_3], [id_4])
) piv
See SQL Fiddle with Demo
But if the values are unknown then you will need to use dynamic SQL:
DECLARE #cols AS NVARCHAR(MAX),
#query AS NVARCHAR(MAX)
select #cols = STUFF((SELECT distinct ','
+ QUOTENAME(rn)
from
(
select 'id_'+cast(row_number() over(partition by address
order by id) as varchar(20)) rn
from yourtable
) src
FOR XML PATH(''), TYPE
).value('.', 'NVARCHAR(MAX)')
,1,1,'')
set #query = 'SELECT address,' + #cols + ' from
(
select address, id,
''id_''+cast(row_number() over(partition by address
order by id) as varchar(20)) rn
from yourtable
) x
pivot
(
max(id)
for rn in (' + #cols + ')
) p '
execute(#query)
See SQL Fiddle with Demo
The result of both queries is:
| ADDRESS | ID_1 | ID_2 | ID_3 | ID_4 |
-------------------------------------------
| 12AnyStreet | 1234 | 1235 | 1236 | 1237 |