Row to column in SQL Server with dynamic number of rows - sql

Question

My source data looks something like this
ID Phone Type Phone Number
-------------------------------------
308820 P 8136542273
308820 F 8136541384
308820 P 8139108555
308820 P 8136842229
308820 F 8139108655
211111 P 6874598695
456788 F 5687659867
In the above data, Phone type = P is phone and F is fax.
I need to sort the data and then pick only one F type and one P type phone number and populate the data as below
ID Fax Number Phone Number
-------------------------------------
308820 8136541384 8136542273
211111 6874598695
456788 5687659867
Can someone help me out how to achieve this. There can be n number of Phone Numbers and Fax numbers against one ID but I need to sort it and pick the first.

Answer 1

You can use conditional aggregation:
select id, max(case when phone_type = 'F' then phone_number end) as fax,
max(case when phone_type = 'P' then phone_number end) as phone
from t
group by id;

Answer 2

Guys thanks for the response. The query which I created is below. The help I got above was used in my query below. However the problem is that in the Binary_CheckSum it randomly picks up a Phone number and as such again selects the same record.
What I am trying to find is that for a particular ID and AddressID combination has the iSActive flag or Phone Number or Fax number has changed so that I need to select that record and insert it in PrescriberLocation table.
select TESTID, ADDRESSID, max(PhoneNumber) as PhoneNumber,max(FaxNumber) as FaxNumber,'1' as isactive from (select distinct a.TESTID
, b.AddressId , CASE WHEN c.PhoneType = 'P'THEN C.PhoneNumber END PhoneNumber
, CASE WHEN c.PhoneType = 'F'THEN C.PhoneNumber END FaxNumber
,'1' as isactive from stg_Address a inner join stg_AddressPhone c on a.TESTID = c.TESTID and a.AddressID = c.AddressID INNER join pbmsys_new.dbo.sc_Address b on
upper(a.Address1) = upper(b.Address1) and upper(isnull(a.Address2,'')) = upper(b.Address2) join pbmsys_new.[dbo].[dr_PrescriberLocation] d
on a.TESTID = d.TESTID and b.AddressId = d.AddressId
where BINARY_CHECKSUM(1,c.PhoneNumber, FaxNumber) != BINARY_CHECKSUM(d.IsActive,d.PhoneNumber,d.FaxNumber) and d.PrescriberLocationId = SELECT max(Z.PrescriberLocationId) as PrescriberLocationId FROM pbmsys_new.dbo.dr_PrescriberLocation Z where d.TESTID = z.TESTID and d.AddressId =z.AddressId))f group by TESTID, AddressId

Answer 3

you can use aggregate function and sub-query
select id,max(fax_Number) as fax_Number, max(phone_Number) as phone_Number
from
(select id, case when phone_type = 'F' then phone_number end as fax_Number,
case when phone_type = 'P' then phone_number end as phone_Number
from yourtable
) as t group by id

Answer 4

If the requirement is to pick first value and not the max value then you can use it.
; with CTE1
As
(
select row_number() over (partition by id,phonetype order by ID) as Num,ID,
case when phonetype = 'F' then PhoneNumber end as FaxNumber,
case when phonetype = 'P' then PhoneNumber end as PhoneNumber
from ContactDetails
)
select ID,max(FaxNumber),max(PhoneNumber)
from (select ID,FaxNumber,PhoneNumber from CTE1 where num = 1)as T2
group by ID