SQL index match to find duplicate data - sql

Question

I have the following table
Code Name Task
aa jones DC
ab dave DC
aca james IF
aca james DC
ab trevor IF
aa jones IF
ag francis DC
ag francis IF
af derek SF
af derek DC
This is a very big table, above is just a quick example.
So, I would like some help finding the code and name that have completed a IF or SF task and a DC task.
I would like it to show where one person has touched both of these tasks. The hierarchy of the tasks is; it comes in as either a SF or IF then someone will do that, then off the back of that we receive a DC task, and I want the ones where it has been completed by the same person, with the same reference number.
I am able to do this in excel with an INDEX MATCH function, but this takes up a tremendous amount of calculation time due to the size of the table.

Answer 1

One way to approach this is using group by with a having. This is a flexible way of expressing these types of conditions:
select code, name
from table t
group by code, name
having sum(case when task = 'DC' then 1 else 0 end) > 0 and
sum(case when task in ('IF', 'SF') then 1 else 0 end) > 0;
Each condition in the having clause counts the number of rows that meet the particular condition. The first, for instance, counts the rows that match 'DC' and takes only the code, name pairs that have at least one such match.

Answer 2

SELECT code,name FROM YOUR_TABLE_NAME WHERE task = 'DC' AND (task = 'IF' OR task = 'SF') GROUP BY name
try this query

Answer 3

Gordon Linoff's query can be made easier under the hypothesys that IF and SF are synonym and cannot be both present for the same Code-Name couple, as the data provided by the OP suggests
SELECT code, name
FROM table t
GROUP BY code, name
HAVING SUM(CASE WHEN task IN ('IF', 'SF', 'DC') THEN 1 ELSE 0 END) = 2;

Answer 4

select code,name from (select distinct code,name from table1 where task='SF' or task='IF') as temp1 inner join (select distinct code as code2,name as name2 from table1 where task='DC') as temp2 on code=code2,name=name2;
I'm assuming that you have the table in table1. The code constructs two tables temp1 and temp2. temp1 contains those codes and names which have been assigned SF and IF. temp2 contains those codes and names which have been assigned DC. Finally, I join the two tables together to find code-name pairs in both tables. This is faster than in Excel because the database engine probably temporarily indexes the columns being joined on.
Actually, you can do this in Excel. You sort the table by code and name, then enter the following formulas (assuming "Code" is in A1):
D2=if(and(A2=A1,B2=B1,D1),true,or(C2="IF",C2="SF"))
E2=if(and(A2=A1,B2=B1,E1),true,C2="DC")
Select these two cells, and double-click the fill-handle (the little square at the bottom right of the selection). Then, with the two columns selected, copy, and then "Paste Special..." > "Values". Then, filter (Alt-D-F-F) for the rows with values in columns D and E being both true. That is the result you want. Select these rows and copy to a new sheet if desired.
Alternatively, you can follow the SQL "group by" solution given by Gordon, so that you do not need to sort: Create two new columns like the above, but:
D1: "D"
E1: "E"
D2=if(or(C2="IF",C2="SF"),1,0)
E2=if(C2="DC",1,0)
Then, "Insert" > "PivotTable", drag "Code" and "Name" to be row labels. Drag "D" to be under Values, click on it, "Value Field Settings...", and then select "Max". Do the same for "E", and then the rows with 1 in both D and E will be the result you want.