SQL join into a Recursive CTE with parameters - sql-server

Question

I am trying to use an SQL Query to create some client side reporting for my company. There exists 3 tables that I would like to join together. One of the tables may require a CTE as I need to recursively go through a table and return a row. Here is how the tables are structured (simply).
I want a output table that, for each WorkOrder, displays the most recently completed task in DataCollection (including its time) and the next Op in the TaskListing. I figured a CTE maybe is the only way to recursively go through each row and determine what task is next. (By checking if the completed Op exists in PreOp column). If the completed cell doesn't exist as a preOp it should default to the MAX(Op) (the last task).
CREATE TABLE [dbo].[WorkOrder](
[WorkOrderID][int] NOT NULL PRIMARY KEY,
[Column1] [nvarchar](20),
[Column2] [nvarchar](20)
)
INSERT INTO WorkOrder VALUES(1,'x','y');
INSERT INTO WorkOrder VALUES(2,'x','y');
INSERT INTO WorkOrder VALUES(3,'x2','y2');
CREATE TABLE [dbo].[DataCollection](
[DataCollection][int] NOT NULL PRIMARY KEY,
[WorkOrderID][int] NOT NULL FOREIGN KEY REFERENCES WorkOrder(WorkOrderID),
[CellTask] [nvarchar](20),
[TimeCompleted] [DateTime]
)
INSERT INTO DataCollection VALUES(1,1,'cella','2016-08-09 00:00:00');
INSERT INTO DataCollection VALUES(2,1,'cellb','2016-08-10 00:00:00');
INSERT INTO DataCollection VALUES(3,1,'cellc','2016-08-11 00:00:00');
INSERT INTO DataCollection VALUES(4,2,'cella','2016-08-09 00:00:00');
INSERT INTO DataCollection VALUES(5,2,'cellb','2016-08-10 00:00:00');
CREATE TABLE [dbo].[TaskListing](
[TaskListingID][int] NOT NULL PRIMARY KEY,
[WorkOrderID][int] NOT NULL FOREIGN KEY REFERENCES WorkOrder(WorkOrderID),
[Op][nvarchar](20) NOT NULL,
[preOP][nvarchar](20),
[CellTask][nvarchar](20) NOT NULL,
[Completed][bit] NOT NULL
)
INSERT INTO TaskListing VALUES(1,1,'10',NULL,'cella',0);
INSERT INTO TaskListing VALUES(2,1,'20','10','cellb',0);
INSERT INTO TaskListing VALUES(3,1,'30',NULL,'cellc',1);
INSERT INTO TaskListing VALUES(4,1,'40','10,30','celld',0);
INSERT INTO TaskListing VALUES(5,2,'10',NULL,'cella',1);
INSERT INTO TaskListing VALUES(6,2,'20','10','cellb',1);
INSERT INTO TaskListing VALUES(7,2,'30','20','cellc',0);
The Output table will represent, for each WorkOrder, the most recently completed cell (from the DataCollection Table, TimeCompleted column) & The next cell in the Work Flow (by looking at the rows on the TaskListing Table for the given WorkOrderID and looking for a row that contain the completed task as a 'PreOp'). If it can't find the completed task as a preOp for any other row it should default to the last task.
The part of the Query I'm having most trouble with is filling in the NextTaskCell column. I need to write a query that looks at all the tasks for a given WorkOrderID (In the TaskListing Table) and based on the completed task, determine what is the next task. I'm finding it difficult to feed in both a WorkOrderID & CellTask then find an instance of itself in the PreOp column.
Output Table
+-------------+-------------------+---------------------+--------------+
| WorkOrderId | LastCompletedCell | CompletedOn | NextTaskCell |
|(WorkOrder) | (DataCollection) | (DataCollection) |(TaskListing) |
+-------------+-------------------+---------------------+--------------+
| 1 | cellc | 2016-08-11 00:00:00 | celld |
| 2 | cellb | 2016-08-10 00:00:00 | cellc |
+-------------+-------------------+---------------------+--------------+
I thank you in advance for your time. If there is any other questions please let me know and I'll try to answer them.
Link to SQL Fiddle SQL Fiddle

Answer 1

The following query gives you the expected output you have in your question. You should test this query against a larger dataset to make sure it is correct in all cases.
;WITH
mtc AS ( -- most recent completion date/time for a work order
SELECT
dc.WorkOrderID,
TimeCompleted=MAX(dc.TimeCompleted)
FROM
DataCollection AS dc
GROUP BY
dc.WorkOrderID
),
lop AS ( -- last operation for work order
SELECT
tl.WorkOrderID,
LastOp=MAX(CAST(tl.Op AS INT))
FROM
TaskListing AS tl
GROUP BY
tl.WorkOrderID
)
SELECT
mtc.WorkOrderID,
LastCompletedCell=dc.CellTask,
CompletedOn=dc.TimeCompleted,
NextTaskCell=ISNULL(tl_next.CellTask,tl_last.CellTask)
FROM
mtc
INNER JOIN DataCollection AS dc ON -- the last completed CellTask
dc.WorkOrderID=mtc.WorkOrderID AND
dc.TimeCompleted=mtc.TimeCompleted
INNER JOIN TaskListing AS tl ON -- Op for CellTask
tl.WorkOrderID=mtc.WorkOrderID AND
tl.CellTask=dc.CellTask
INNER JOIN lop ON
lop.WorkOrderID=mtc.WorkOrderID
INNER JOIN TaskListing AS tl_last ON -- CellTask for last Op
tl_last.WorkOrderID=mtc.WorkOrderID AND
tl_last.Op=lop.LastOp
LEFT JOIN TaskListing AS tl_next ON -- Look for next CellTask where Op is a PreOp of another CellTask
tl_next.WorkOrderID=mtc.WorkOrderID AND
','+tl_next.preOP+',' LIKE '%,'+tl.Op+',%'
ORDER BY
mtc.WorkOrderId;
Note: It is a bad idea to store PreOps as a comma-separated string. This is not how you should store data in relational databases. When you do, you will have to resort to more complex and less efficient queries. To wit, see the join condition in tl_next.
Instead you should have a table to store PreOps as separate rows, linking to the parent Op that depends on it.