SQL Backfilling Dates in a Recruiting Funnel - sql

Question

I am working with recruiting funnel data but a lot of the data has missing date stamps for some upstream stages of the funnel (recruiters either skipping or not properly recording candidate movements through the funnel)
I can fetch data with the steps and their order, my objective is to backfill missing entry dates that are upstream from a stage with an entry date (I figured the backfill could be the same date as the date for the next stage that has a date)
Any ideas on how I can do this using SQL?
I created some sample data:
CREATE TABLE stages (
job_id INT,
application_id INT,
stage_name VARCHAR,
stage_order INT,
entered_on DATE,
exited_on DATE
)
;
INSERT INTO stages (job_id, application_id, stage_name, stage_order, entered_on, exited_on)
VALUES (8339915, 24342, 'Application Review', 0, '2015-06-06', '2015-06-22'),
(8339915, 24342, 'Hiring Manager Review', 1, NULL, NULL),
(8339915, 24342, 'Recruiter Phone Screen', 2, '2015-06-07', '2015-06-22'),
(8339915, 24342, 'Phone Interview', 3, NULL, NULL),
(8339915, 24342, 'Phone Interview 2', 4, '2015-06-22', '2015-07-06'),
(8339915, 24342, 'Face to Face', 5, '2015-07-06', '2015-07-24'),
(8339915, 24342, 'Face to Face 2', 6, NULL, NULL),
(8339915, 24342, 'Offer', 7, NULL, NULL),
(8339915, 24342, 'Hired', 1000, NULL, NULL)
;
SELECT *
FROM stages
ORDER BY job_id ASC, application_id ASC, stage_order

Answer 1

This code would help with your requirement but the intermediate join is painful, if you have a low amount of stages this would work just fine. let me know if there is any concern.
Select A.job_id,A.application_id,a.stage_name,a.stage_order,isnull(a.entered_on,ent.entered_on),isnull(a.exited_on,exi.exited_on)
from stages A
left join
(select * from(select *, row_number() over(partition by job_id,application_id,s1 order by s1,s2 desc) rn from(
select job_id,application_id,stage_name,a.stage_order as s1,b.stage_order as s2,isnull(a.entered_on,b.entered_on) as entered_on from stages A
inner join (select distinct stage_order,entered_on from stages where entered_on is not null
)b on A.stage_order>=b.stage_order
where a.entered_on is null) j)k
where K.rn=1)Ent on Ent.job_id=a.job_id and ent.application_id=a.application_id and ent.s1=a.stage_order
left join(
select * from(select *, row_number() over(partition by job_id,application_id,s1 order by s1,s2 desc) rn from(
select job_id,application_id,stage_name,a.stage_order as s1,b.stage_order as s2,isnull(a.exited_on,b.exited_on) as exited_on from stages A
inner join (select distinct stage_order,exited_on from stages where exited_on is not null
)b on A.stage_order>=b.stage_order
where a.exited_on is null) j)k
where K.rn=1)Exi on Exi.job_id=a.job_id and exi.application_id=a.application_id and exi.s1=a.stage_order