This is not really the create table and insert statements I asked for and pointed you to in the article. since you brand new around here I cobbled this together based on what you did post so you can see what I mean.

if OBJECT_ID('tempdb..#Sales') is not null

drop table #Sales

create table #Sales

(

SalesOrder char(6),

Invoice varchar(10),

InvoiceDate datetime,

InvoiceAmount int

)

insert #Sales

select 'SO-100', 'INV-38222', '4/12/2013', 7 union all

select 'SO-100', 'INV-38222', '4/12/2013', 4 union all

select 'SO-100', 'INV-38222', '4/12/2013', 4 union all

select 'SO-100', 'INV-38222', '4/12/2013', 7 union all

select 'SO-101', 'INV-4825', '4/14/2013', 10 union all

select 'SO-101', 'INV-4825', '4/14/2013', 8 union all

select 'SO-101', 'INV-4827', '4/15/2013', 3 union all

select 'SO-102', 'INV-4830', '4/20/2013', 8 union all

select 'SO-102', 'INV-4830', '4/20/2013', 8 union all

select 'SO-102', 'INV-4830', '4/20/2013', 8 union all

select 'SO-102', 'INV-4831', '4/21/2013', 63 union all

select 'SO-102', 'INV-4832', '4/22/2013', 26

I know this doesn't seem like it is so hard but if you post the ddl and sample data there are two major advantages for you. First, we don't have to guess at column names and more importantly datatypes. The second advantage is that since everyone who posts here is a volunteer, if you post everything in a readily consumable format you are much more likely to get more people helping you.

So now to pull this data like you stated, it is nothing more than a group by. Please note that some of the SalesOrders will show up multiple times because they were invoiced and shipped on different days. And since you want the invoice date in the output you either show the values or you have to aggregate them (MIN, MAX, etc)

select SalesOrder, InvoiceDate, count(Invoice) as NumShipments

from #Sales

group by SalesOrder, InvoiceDate