Find all rows where start date is before a prior end date - sql

Question

Is there a way to pull back all records that have overlapping datetimes based on a user?
For instance;
TableA has the following rows;
TrainerID StartTime EndTime
1234 10-1-2015 08:30 10-1-2015 09:00
1234 10-1-2015 08:45 10-1-2015 09:15
1234 10-1-2015 09:30 10-1-2015 10:00
2345 10-1-2015 08:45 10-1-2015 09:15
2345 10-1-2015 09:30 10-1-2015 10:00
I need a query that can pull ONLY the following record because it's start time is before the previous end time for the trainer (double booked):
1234 10-1-2015 08:45 10-1-2015 09:15

Answer 1

The EXIST code below should give you that answer. The code ensures that the start time of the clashing entry is before the start of the main list entry while the start time of the clash is still after the start time of the mail list entry.
SELECT *
FROM tblTest clashing
WHERE EXISTS
(
SELECT 1
FROM tblTest mainlist
WHERE clashing.trainderid = mainlist.trainderid
AND clashing.starttime < mainlist.endtime
AND clashing.starttime > mainlist.starttime
)
This can also be written with an IN statement, but EXIST is much more efficient

Answer 2

To remove overlapping dates you can use:
Demo
CREATE TABLE #TABLEA( TrainerID INT, StartDate DATETIME, EndDate DATETIME);
INSERT INTO #TABLEA
SELECT 1234, '10-1-2015 08:30', '10-1-2015 09:00'
UNION ALL SELECT 1234 , '10-1-2015 08:45', '10-1-2015 09:15'
UNION ALL SELECT 1234 , '10-1-2015 09:30', '10-1-2015 10:00'
UNION ALL SELECT 2345 , '10-1-2015 08:45', '10-1-2015 09:15'
UNION ALL SELECT 2345 , '10-1-2015 09:30', '10-1-2015 10:00';
SELECT
D.TrainerID,
[StartTime] = D.StartDate,
[EndTime] = (SELECT MIN(E.EndDate)
FROM #TABLEA E
WHERE E.EndDate >= D.EndDate
AND E.TrainerID = D.TrainerID
AND NOT EXISTS (SELECT 1
FROM #TABLEA E2
WHERE E.StartDate < E2.StartDate
AND E.EndDate > E2.StartDate
AND E.TrainerID = E2.TrainerID))
FROM #TABLEA D
WHERE NOT EXISTS ( SELECT 1
FROM #TABLEA D2
WHERE D.StartDate < D2.EndDate
AND D.EndDate > D2.EndDate
AND D.TrainerID = D2.TrainerID);

Answer 3

You can use below code to get the required row, however based on your logic row from next trainer id (i.e. 2345) will also be qualified
DECLARE #Trainers TABLE
(
TrainerId INT,
Start_Time datetime,
End_Time datetime
)
INSERT INTO #Trainers VALUES
(1234,'10-1-2015 08:30','10-1-2015 09:00 '),
(1234,'10-1-2015 08:45','10-1-2015 09:15'),
(1234,'10-1-2015 09:30','10-1-2015 10:00'),
(2345 ,' 10-1-2015 08:45','10-1-2015 09:15'),
(2345 ,' 10-1-2015 09:30 ',' 10-1-2015 10:00')
;WITH TrainersTemp AS
(
SELECT *, ROW_NUMBER() OVER ( ORDER BY trainerid) AS rn
FROM #Trainers
)
SELECT CX.TrainerId, CX.Start_Time, CX.End_Time
FROM TrainersTemp CX JOIN TrainersTemp CY
ON CX.rn = CY.rn + 1
WHERE CY.End_Time < CX.Start_Time
Demo (SQL fiddle is down again)
or if you want to see all rows except the faulty one then use below code
;WITH TrainersTempAll AS
(
SELECT *, ROW_NUMBER() OVER ( ORDER BY trainerid) AS rn
FROM #Trainers
)
SELECT CX.TrainerId, CX.Start_Time, CX.End_Time
FROM TrainersTempAll CX JOIN TrainersTempAll CY
ON CX.rn = CY.rn + 1

Answer 4

Firstly you should sort by trainerId and Start_time. And then join two tables with correct condition.
Try this query:
;WITH TrainersTemp AS
(
SELECT *, ROW_NUMBER() OVER ( ORDER BY trainerid, Start_Time) AS row_num
FROM Trainers
)
select t2.* from TrainersTemp t1
join TrainersTemp t2 on t1.TrainerId = t2.TrainerId and t1.row_num = t2.row_num-1
where t2.Start_Time<t1.End_Time

Answer 5

As you use SQL Server 2012 you can use LAG function, which would be likely more efficient than self-join. The query becomes pretty simple as well.
For each row LAG gives you EndTime from the previous row (partitioned by TrainerID). Then just compare StartTime from the current row with EndTime from the previous row.
SQL Fiddle
WITH
CTE
AS
(
SELECT
TrainerID
,StartTime
,EndTime
,LAG(EndTime) OVER(PARTITION BY TrainerID ORDER BY StartTime) AS PrevEndTime
FROM TableA
)
SELECT
TrainerID
,StartTime
,EndTime
FROM CTE
WHERE StartTime < PrevEndTime
;
Results:
| TrainerID | StartTime | EndTime |
|-----------|---------------------------|---------------------------|
| 1234 | October, 01 2015 08:45:00 | October, 01 2015 09:15:00 |