I have to identify the records with check_in category of S1, T1 and P1 in the check-in-order and pivot the rest of the data until we hit the next category. I was able to mark the category but was unsuccessful to pickup the min and max dates of each category as the check_in values are different for every row.
Here is the sample data and the expected results. Appreciate your help.
Source:
GRPNBR CHECK_IN Date_of_check_in CHECK_IN_ORDER
200-001 S1 1/15/2020 3
200-001 Y23 3/4/2020 15
200-001 M56 5/6/2020 17
200-001 UN 8/31/2020 38
200-001 T1 10/12/2020 78
200-001 C T28 11/23/2020 91
200-001 C M3 1/29/2021 93
200-001 P1 4/22/2021 94
200-001 CM9 8/4/2021 95
200-001 CM10 9/4/2021 97
Expected Result:
GRPNBR CHECK_IN POST_CHECK_IN Date_of_check_in CHECK_IN_ORDER
200-001 S1 Y23 3/4/2020 15
200-001 S1 M56 5/6/2020 17
200-001 S1 UN 8/31/2020 38
200-001 T1 C T28 11/23/2020 91
200-001 T1 C M3 1/29/2021 93
200-001 P1 CM9 8/4/2021 95
200-001 P1 CM10 9/4/2021 97
CodePudding user response:
Here is one approach that utilizes the window functions to flag and sequence data and then a simple aggregation for the final results
CodePudding user response:
Your data
GRPNBR | CHECK_IN | Date_of_check_in | CHECK_IN_ORDER | FIELD5 |
---|---|---|---|---|
200-001 | S1 | 1/15/2020 | 3 | |
200-001 | Y23 | 3/4/2020 | 15 | |
200-001 | M56 | 5/6/2020 | 17 | |
200-001 | UN | 8/31/2020 | 38 | |
200-001 | T1 | 10/12/2020 | 78 | |
200-001 | C | T28 | 11/23/2020 | 91 |
200-001 | C | M3 | 1/29/2021 | 93 |
200-001 | P1 | 4/22/2021 | 94 | |
200-001 | CM9 | 8/4/2021 | 95 | |
200-001 | CM10 | 9/4/2021 | 97 |
to get value like S1, T1 and P1 ,use following query
select GRPNBR,CHECK_IN,Date_of_check_in
from YourTable
where CHECK_IN like '_1'
GRPNBR | CHECK_IN | Date_of_check_in |
---|---|---|
200-001 | S1 | 2020-01-15 |
200-001 | T1 | 2020-10-12 |
200-001 | P1 | 2021-04-22 |
then you want to join with following query
select GRPNBR,CHECK_IN as POST_CHECK_IN ,Date_of_check_in,CHECK_IN_ORDER
from YourTable
where CHECK_IN not like '_1'
with following result to get your desired result
GRPNBR | POST_CHECK_IN | Date_of_check_in | CHECK_IN_ORDER |
---|---|---|---|
200-001 | Y23 | 2020-03-04 | 15 |
200-001 | M56 | 2020-05-06 | 17 |
200-001 | UN | 2020-08-31 | 38 |
200-001 | C T28 | 2020-11-23 | 91 |
200-001 | C M3 | 2021-01-29 | 93 |
200-001 | CM9 | 2021-08-04 | 95 |
200-001 | CM10 | 2021-09-04 | 97 |
I used Pivot
, Subquery
and Case
to get your desired result
SELECT A.grpnbr,
CASE
WHEN date_of_check_in > s1
AND date_of_check_in < t1 THEN 'S1'
WHEN date_of_check_in > t1
AND date_of_check_in < p1 THEN 'T1'
WHEN date_of_check_in > p1 THEN 'P1'
END AS CHECK_IN,
post_check_in,
date_of_check_in,
check_in_order
FROM (SELECT grpnbr,
check_in AS POST_CHECK_IN,
date_of_check_in,
check_in_order
FROM yourtable
WHERE check_in NOT LIKE '_1') A
JOIN (SELECT *
FROM (SELECT grpnbr,
check_in,
date_of_check_in
FROM yourtable
WHERE check_in LIKE '_1') src
PIVOT ( Max(date_of_check_in)
FOR check_in IN ([S1],
[T1],
[P1]) ) piv) B
ON A.grpnbr = B.grpnbr