CODEWITHSUNDEEP
sqlsql-servert-sql
Kric
Kric

Reputation: 103

Several rows with exact same values but only one column returns different value; How to grab this value and create another column(s) for that?

So Here it is : I joined several tables and get results like this:

ContactID    Name    Title      Phone
1234567      John    Manager    999999
1234567      John    Manager    888888

I want to get the unique ContactID in my result and put another phone value to 'Phone2'. (There are more than 2, Phone3, Phone4)

So the result I want to see is like this

ContactID    Name   Title             Phone    Phone2     Phone3    Phone4...
1234567      John   Manager           999999    888888     NULL      NULL

Any help? I tried to use Pivot, but it seems like the aggregate function wont work in this scenario.

Upvotes: 0

Views: 2164

Answers (2)

RichardTheKiwi
RichardTheKiwi

Reputation: 107786

You can still use a PIVOT, although you have to make up a pivot column, for which I use the ROW_NUMBER() function below. I have allowed for 4 phone numbers below, just expand the pattern, e.g. there are 2 places with [4], if you need to cater for more.

Below I show data coming from a sample table named "data". If you have a complex query, you can first name it using a Common Table Expression, i.e.

;with data as (
   <your entire complex query that produces the
    columns: ContactId, Name, Title, Phone>
)
select ContactID, Name, Title,
       [1] as Phone1,
       [2] as Phone2,
       [3] as Phone3,
       [4] as Phone4
  from (
    select ContactID, Name, Title, Phone,
           Row_Number() over (partition by ContactID, Name, Title
                              order by Phone) RN
      from data) p
pivot (max(Phone) for RN in ([1],[2],[3],[4])) v;

SQL Fiddle

MS SQL Server 2012 Schema Setup:

create table data (
  ContactID int,
  Name varchar(10),
  Title varchar(10),
  Phone varchar(20));
insert data values
(1234561,      'Jill',    'Employee',    999991),
(1234562,      'Joel',    'Employee',    999992),
(1234567,      'John',    'Manager',    999999),
(1234567,      'John',    'Manager',    888888);

Query 1:

select ContactID, Name, Title,
       [1] as Phone1,
       [2] as Phone2,
       [3] as Phone3,
       [4] as Phone4
  from (
    select ContactID, Name, Title, Phone,
           Row_Number() over (partition by ContactID, Name, Title
                              order by Phone) RN
      from data) p
pivot (max(Phone) for RN in ([1],[2],[3],[4])) v;

Results:

| CONTACTID | NAME |    TITLE | PHONE1 | PHONE2 | PHONE3 | PHONE4 |
-------------------------------------------------------------------
|   1234561 | Jill | Employee | 999991 | (null) | (null) | (null) |
|   1234562 | Joel | Employee | 999992 | (null) | (null) | (null) |
|   1234567 | John |  Manager | 888888 | 999999 | (null) | (null) |

Upvotes: 1

Scott Shipp
Scott Shipp

Reputation: 2321

If you want you can get all the phone numbers into one column which might be more useful for you depending on your situation.

First     Last        Phone
John      Manager     888888,999999,111111

See http://www.centrolutions.com/blog/post/2010/03/29/Selecting-Multiple-Rows-as-a-Single-Column-in-SQL.aspx

Upvotes: 0

Related Questions