I have following SQL Server table ITEM:
+------------+-----------+------+--------+-----------+------------+
| Date | item_code | name | in/out | total_qty | SortNumber |
+------------+-----------+------+--------+-----------+------------+
| 08/07/2019 | 001 | A | -50 | 100 | 8 |
| 07/07/2019 | 001 | A | 50 | 100 | 7 |
| 06/07/2019 | 003 | C | 25 | 25 | 6 |
| 05/07/2019 | 001 | A | 50 | 50 | 5 |
| 04/07/2019 | 002 | B | 100 | 200 | 4 |
| 03/07/2019 | 003 | C | -25 | 0 | 3 |
| 02/07/2019 | 003 | C | 25 | 25 | 2 |
| 01/07/2019 | 002 | B | 100 | 100 | 1 |
+------------+-----------+------+--------+-----------+------------+
I've tried:
select itemcode, max(Sort_Number)
from ITEM
group by item_code
order by item_code asc
but I want result:
+---------------------+-----------+------------------+
| Distinct(item_code) | Total_qty | Max(Sort_Number) |
+---------------------+-----------+------------------+
| 001 | 100 | 8 |
| 002 | 200 | 4 |
| 003 | 25 | 6 |
+---------------------+-----------+------------------+
Can anyone help me?
The below query gives you the desired result -
With cteItem as
(
select item_code, total_qty, SortNumber,
Row_Number() over (partition by item_code order by SortNumber desc) maxSortNumber
from ITEM
)
select item_code, total_qty, SortNumber from cteItem where maxSortNumber = 1
just need to add max(sort_number) to your query
select item_code ,max(total_qty), max(sort_number)
from ITEM
group by item_code
order by item_code asc
Hard to phrase the title for this one.
I have a table of data which contains a row per invoice. For example:
| Invoice ID | Customer Key | Date | Value | Something |
| ---------- | ------------ | ---------- | ------| --------- |
| 1 | A | 08/02/2019 | 100 | 1 |
| 2 | B | 07/02/2019 | 14 | 0 |
| 3 | A | 06/02/2019 | 234 | 1 |
| 4 | A | 05/02/2019 | 74 | 1 |
| 5 | B | 04/02/2019 | 11 | 1 |
| 6 | A | 03/02/2019 | 12 | 0 |
I need to add another column that counts the number of previous rows per CustomerKey, but only if "Something" is equal to 1, so that it returns this:
| Invoice ID | Customer Key | Date | Value | Something | Count |
| ---------- | ------------ | ---------- | ------| --------- | ----- |
| 1 | A | 08/02/2019 | 100 | 1 | 2 |
| 2 | B | 07/02/2019 | 14 | 0 | 1 |
| 3 | A | 06/02/2019 | 234 | 1 | 1 |
| 4 | A | 05/02/2019 | 74 | 1 | 0 |
| 5 | B | 04/02/2019 | 11 | 1 | 0 |
| 6 | A | 03/02/2019 | 12 | 0 | 0 |
I know I can do this using either a CTE like this...
(
select
count(*)
from table
where
[Customer Key] = t.[Customer Key]
and [Date] < t.[Date]
and Something = 1
)
But I have a lot of data and that's pretty slow. I know I can also use cross apply to achieve the same thing, but as far as I can tell that's not any better performing than just using a CTE.
So; is there a more efficient means of achieving this, or do I just suck it up?
EDIT: I originally posted this without the requirement that only rows where Something = 1 are counted. Mea culpa - I asked it in a hurry. Unfortunately I think that this means I can't use row_number() over (partition by [Customer Key])
Assuming you're using SQL Server 2012+ you can use Window Functions:
COUNT(CASE WHEN Something = 1 THEN CustomerKey END) OVER (PARTITION BY CustomerKey ORDER BY [Date]
ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) -1 AS [Count]
Old answer before new required logic:
COUNT(CustomerKey) OVER (PARTITION BY CustomerKey ORDER BY [Date]
ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) -1 AS [Count]
If you're not using 2012 an alternative is to use ROW_NUMBER
ROW_NUMBER() OVER (PARTITION BY CustomerKey ORDER BY [Date]) - 1 AS Count
I would like to add a column indicating the number invites a person received before they accepted by incrementally counting the number of null columns before a non-null while partitioning over the PERSON_ID and ordering by the INVITED_DATE.
My table has the following format:
| UNIQUE_ID | PERSON_ID | INVITED_DATE | ACCEPTED_DATE |
| 12345 | 567 | 12-01-18 | NULL |
| 12346 | 567 | 12-02-18 | NULL |
| 12347 | 567 | 12-03-18 | NULL |
| 12348 | 567 | 12-04-18 | 12-04-18 |
| 12349 | 567 | 12-05-18 | NULL |
| 12350 | 568 | 12-01-18 | NULL |
| 12351 | 568 | 12-02-18 | 12-02-18 |
The output should ideally look like the following:
| UNIQUE_ID | PERSON_ID | INVITED_DATE | ACCEPTED_DATE | INVITES_BEFORE_ACCEPT |
| 12345 | 567 | 12-01-18 | NULL | 1 |
| 12346 | 567 | 12-02-18 | NULL | 2 |
| 12347 | 567 | 12-03-18 | NULL | 3 |
| 12348 | 567 | 12-04-18 | 12-04-18 | 0 |
| 12349 | 567 | 12-05-18 | NULL | 1 |
| 12350 | 568 | 12-01-18 | NULL | 1 |
| 12351 | 568 | 12-02-18 | 12-02-18 | 0 |
So far I've tried a number iterations of ROW NUMBER with OVER and PARTITION but I've found it will need to be an OUTER APPLY. The following OUTER APPLY counts over the data but doesn't restart the count with a successful accept.
SELECT t.* , invites.INVITES_BEFORE_ACCEPT
FROM table t
OUTER APPLY (
SELECT COUNT(*) INVITES_BEFORE_ACCEPT
FROM table t2
WHERE t.PERSON_ID = t2.PERSON_ID and t.INVITED_DATE < t2.ACCEPTED_DATE
) invites
One way would be
WITH t
AS (SELECT *,
COUNT(ACCEPTED_DATE)
OVER (
PARTITION BY PERSON_ID
ORDER BY INVITED_DATE) AS Grp
FROM [table])
SELECT *,
SUM(CASE
WHEN ACCEPTED_DATE IS NULL
THEN 1
ELSE 0
END)
OVER (
PARTITION BY PERSON_ID, Grp
ORDER BY INVITED_DATE) AS INVITES_BEFORE_ACCEPT
FROM t
Demo
Let's say I have the following table (data is completely fiction):
ID | MonthDate | PersonID | Name | Status | MonthsAgoSinceLastCheck
1 | 2017-12 | 900 | Jack | Ill | -
2 | 2018-01 | 900 | Jack | Ill | 1
3 | 2018-02 | 900 | Jack | Ill | 2
4 | 2018-03 | 900 | Jack | Healthy | 1
5 | 2017-02 | 901 | Bill | Ill | -
6 | 2017-03 | 901 | Bill | Ill | 1
7 | 2017-05 | 901 | Bill | Healthy | 1
For each record, I would like to see the previous status that person had X months ago since last check (column MonthsAgoSinceLastCheck). Notice that MonthDate can skip months.
So in this case, the result would be
ID | MonthDate | PersonID | Name | Status | MonthsAgoSinceLastCheck | PreviousSatus
1 | 2017-12 | 900 | Jack | Ill | - | -
2 | 2018-01 | 900 | Jack | Ill | 1 | Ill
3 | 2018-02 | 900 | Jack | Ill | 2 | Ill
4 | 2018-03 | 900 | Jack | Healthy | 1 | Ill
5 | 2017-02 | 901 | Bill | Healthy | - | -
6 | 2017-03 | 901 | Bill | Healthy | 1 | Healthy
7 | 2017-05 | 901 | Bill | Ill | 2 | Healthy
Any sugestions/tips? I tried to do this with CTE's and self-joins but failed on both.
It's way easier to use full dates than year and months separately. The first thing you should do is generate a full date from your year + month. Then just self join with previous month, depending on the last check.
;WITH DataWithDates AS
(
SELECT
T.ID,
MonthDate = CONVERT(DATE, T.MonthDate + '-01'),
T.PersonID,
T.Name,
T.Status,
T.MonthsAgoSinceLastCheck
FROM
YourTable AS T
)
SELECT
D.ID,
D.MonthDate,
D.PersonID,
D.Name,
D.Status,
D.MonthsAgoSinceLastCheck,
PreviousStatus = N.Status
FROM
DataWithDates AS D
LEFT JOIN DataWithDates AS N ON
D.PersonID = N.PersonID AND
N.MonthDate = DATEADD(MONTH, -1 * D.MonthsAgoSinceLastCheck, D.MonthDate)
I'm assuming your MonthDate has values for all rows, otherwise the conversion will fail. I'm also assuming that your - values for MonthsAgoSinceLastCheck are actually NULL.
try this:
select *,LAG(Status) OVER(Partition by Name Order by MonthDate,Id) AS PreviousSatus
from tab1
order by id
SQl Fiddle:http://sqlfiddle.com/#!18/04407/4
I have three SQL tables: Release (which represents a release of a movie), Media (which represents the individual pieces of recordable media in those releases; i.e. for Blu-ray/DVD combos, there will be two rows in Media, one Blu-ray and one DVD, that point back to the same row in Release) and MediaType (which defines Blu-ray, DVD, VHS, etc.). There's a one-to-many relationship for Release/Media and MediaType/Media, with Media being on the "many" side of both relationships. I have a view for Release, vRelease, which contains aggregate functions, such as a COUNT that shows how many media are associated with that release. This is what I have for this view so far:
SELECT dbo.Release.ReleaseID
,dbo.Release.Name
,CASE WHEN Release.Compilation = 0 THEN 'No' WHEN Release.Compilation = 1 THEN 'Yes' END AS Compilation
,dbo.Release.Owner
,CASE WHEN Release.LentOut = 0 THEN 'No' WHEN Release.LentOut = 1 THEN 'Yes' END AS LentOut
,COUNT(dbo.Media.ReleaseID) AS NumberOfMedia
,MIN(dbo.Media.MediaID) AS FirstMediaID
,MIN(dbo.MediaType.Name) AS FirstMediaType
FROM dbo.MediaType INNER JOIN
dbo.Media ON dbo.MediaType.MediaTypeID = dbo.Media.MediaTypeID RIGHT OUTER JOIN
dbo.Release ON dbo.Media.ReleaseID = dbo.Release.ReleaseID
GROUP BY dbo.Release.ReleaseID, dbo.Release.Name, dbo.Release.Compilation, dbo.Release.Owner, dbo.Release.LentOut
You'll notice that I've also included two other aggregate columns: FirstMediaID grabs the ID of the media associated with that release that appears first in the Media table (i.e. if a release has two DVDs associated with it, it gets one with the lower ID value). This column on its own isn't useful; what I want to do is then, in turn, get the MediaType that that Media is associated with. In other words, I want a column that shows the MediaType of the first Media that is attached to each Release. The column after that, FirstMediaType, is supposed to do that, but it instead gets the MediaType among all of the Media associated with the Release and picks the one that is alphabetically first - which means that Blu-ray will always be prioritized over DVD (which is fine), but Audio CD will always be prioritized over everything else (which is not fine).
How do I get the FirstMediaType column in this view to get the MediaType of the Media identified in FirstMediaID?
UPDATE: Here are the tables, their columns and some sample rows.
A couple from Release:
+-----------+----------------------------------------+-------+-------------+---------+
| ReleaseID | Name | Owner | Compilation | LentOut |
+-----------+----------------------------------------+-------+-------------+---------+
| 2 | Alice in Wonderland | NULL | 0 | 0 |
| 6 | 4 Film Favorites - Family Comedies | NULL | 1 | 0 |
| 8 | Aladdin | NULL | 0 | 0 |
| 463 | Harry Potter and the Half-Blood Prince | NULL | 0 | 1 |
| 534 | Spirited Away | Ryan | 0 | 0 |
| 571 | The Original Christmas Classics | NULL | 1 | 0 |
+-----------+----------------------------------------+-------+-------------+---------+
Compilation indicates a release that has more than one movie in it.
Corresponding entries in Media:
+---------+-------------+-------------------------------------------------------------------------------------+-----------+
| MediaID | MediaTypeID | Name | ReleaseID |
+---------+-------------+-------------------------------------------------------------------------------------+-----------+
| 2 | 2 | Movie | 2 |
| 3 | 1 | Movie | 2 |
| 12 | 1 | Space Jam; Looney Tunes: Back in Action | 6 |
| 13 | 1 | Funky Monkey; Osmosis Jones | 6 |
| 17 | 3 | Movie | 8 |
| 620 | 1 | Movie | 463 |
| 726 | 1 | Movie | 534 |
| 807 | 1 | Rudolph the Red-Nosed Reindeer; Cricket on the Hearth | 571 |
| 808 | 1 | Frosty the Snowman; Frosty Returns | 571 |
| 809 | 1 | Santa Claus is Comin' to Town!; Mr. Magoo's Christmas Carol; The Little Drummer Boy | 571 |
| 810 | 4 | Tracks 1-7 | 571 |
+---------+-------------+-------------------------------------------------------------------------------------+-----------+
First few in MediaType:
+-------------+--------------+
| MediaTypeID | Name |
+-------------+--------------+
| 1 | DVD Disc |
| 2 | Blu-ray Disc |
| 3 | VHS |
| 4 | Audio CD |
+-------------+--------------+
The corresponding entries in vRelease SHOULD be this:
+-----------+----------------------------------------+-------------+-------+---------+---------------+--------------+----------------+
| ReleaseID | Name | Compilation | Owner | LentOut | NumberOfMedia | FirstMediaID | FirstMediaType |
+-----------+----------------------------------------+-------------+-------+---------+---------------+--------------+----------------+
| 2 | Alice in Wonderland | No | NULL | No | 2 | 2 | Blu-ray Disc |
| 6 | 4 Film Favorites - Family Comedies | Yes | NULL | No | 2 | 12 | DVD Disc |
| 8 | Aladdin | No | NULL | No | 1 | 17 | VHS |
| 463 | Harry Potter and the Half-Blood Prince | No | NULL | Yes | 1 | 620 | DVD Disc |
| 534 | Spirited Away | No | Ryan | No | 1 | 726 | DVD Disc |
| 571 | The Original Christmas Classics | Yes | NULL | No | 4 | 807 | DVD Disc |
+-----------+----------------------------------------+-------------+-------+---------+---------------+--------------+----------------+
But it's actually this:
+-----------+----------------------------------------+-------------+-------+---------+---------------+--------------+----------------+
| ReleaseID | Name | Compilation | Owner | LentOut | NumberOfMedia | FirstMediaID | FirstMediaType |
+-----------+----------------------------------------+-------------+-------+---------+---------------+--------------+----------------+
| 2 | Alice in Wonderland | No | NULL | No | 2 | 2 | Blu-ray Disc |
| 6 | 4 Film Favorites - Family Comedies | Yes | NULL | No | 2 | 12 | DVD Disc |
| 8 | Aladdin | No | NULL | No | 1 | 17 | VHS |
| 463 | Harry Potter and the Half-Blood Prince | No | NULL | Yes | 1 | 620 | DVD Disc |
| 534 | Spirited Away | No | Ryan | No | 1 | 726 | DVD Disc |
| 571 | The Original Christmas Classics | Yes | NULL | No | 4 | 807 | Audio CD |
+-----------+----------------------------------------+-------------+-------+---------+---------------+--------------+----------------+
It's that last one that's the problem.
I ended up finding a simple way to do what I wanted. It isn't as fancy as Used_By_Already's answer (which did end up working, as far as I could tell) and probably breaks a SQL Best Practices rule somewhere, but it's much easier to understand and maintain - at least for my newbie brain.
Since the problem was trying to get the view to use an aggregate column it calculated in a join, I just split the two-step action over two views. vReleasePre has all of the columns I outlined in my original query except for FirstMediaType. vRelease now simply takes all of the columns from vReleasePre and adds FirstMediaType, which takes its value from a join at the end: LEFT OUTER JOIN dbo.vMedia ON dbo.vReleasePre.FirstMediaID = dbo.vMedia.MediaID, where vMedia is a view with all the columns from Media, plus the MediaType column (I already had vMedia lying around).
Since this database is being used in an ASP.NET MVC web application via Entity Framework, and EF has been pretty strange about what it will and won't accept into the data model, I figure that a simple, if roundabout, solution is probably going to be my best option.
vReleasePre:
SELECT dbo.Release.ReleaseID
,dbo.Release.Name
,CASE WHEN Release.Compilation = 0 THEN 'No' WHEN Release.Compilation = 1 THEN 'Yes' END AS Compilation
,dbo.Release.Owner
,CASE WHEN Release.LentOut = 0 THEN 'No' WHEN Release.LentOut = 1 THEN 'Yes' END AS LentOut
,COUNT(dbo.Media.ReleaseID) AS NumberOfMedia
,MIN(dbo.Media.MediaID) AS FirstMediaID
FROM dbo.MediaType INNER JOIN
dbo.Media ON dbo.MediaType.MediaTypeID = dbo.Media.MediaTypeID RIGHT OUTER JOIN
dbo.Release ON dbo.Media.ReleaseID = dbo.Release.ReleaseID
GROUP BY dbo.Release.ReleaseID, dbo.Release.Name, dbo.Release.Compilation, dbo.Release.Owner, dbo.Release.LentOut
vRelease:
SELECT dbo.vReleasePre.ReleaseID
,dbo.vReleasePre.Name
,dbo.vReleasePre.Compilation
,dbo.vReleasePre.Owner
,dbo.vReleasePre.LentOut
,dbo.vReleasePre.NumberOfMedia
,dbo.vMedia.MediaType
FROM dbo.vReleasePre LEFT OUTER JOIN
dbo.vMedia ON dbo.vReleasePre.FirstMediaID = dbo.vMedia.MediaID
A very convenient technique that returns whole rows associated with needs such as "First", "Last", "Earliest", "Latest" is to use row_number() over(). Here you want the "first media type", so it is relevant here.
As you will see in the following query joining the [Media] table is replaced with a subquery that includes a row number calculation. Here we partition by ReleaseID and order by MediaID, so, for each ReleaseID the first row will be the one with the lowest MediaID value. Then in the join to this derived table an extra condition is added to only consider rows with a row number of 1.
Proposed Query
SELECT
r.ReleaseID
, m.MediaID
, mt.MediaTypeID
, mt.name MediaName
, r.Name
, CASE
WHEN r.Compilation = 0 THEN 'No'
WHEN r.Compilation = 1 THEN 'Yes'
END AS compilation
, r.Owner
, CASE
WHEN r.LentOut = 0 THEN 'No'
WHEN r.LentOut = 1 THEN 'Yes'
END AS lentout
FROM dbo.Release r
INNER JOIN (
SELECT
Media.*
, ROW_NUMBER() OVER(PARTITION BY ReleaseID
ORDER BY MediaID) AS rn
FROM dbo.Media
) m ON r.ReleaseID = m.ReleaseID and rn = 1
INNER JOIN dbo.MediaType mt ON m.MediaTypeID = mt.MediaTypeID
Result
| ReleaseID | MediaID | MediaTypeID | MediaName | Name | compilation | Owner | lentout |
|-----------|---------|-------------|--------------|----------------------------------------|-------------|--------|---------|
| 2 | 2 | 2 | Blu-ray Disc | Alice in Wonderland | No | (null) | No |
| 6 | 12 | 1 | DVD Disc | 4 Film Favorites - Family Comedies | Yes | (null) | No |
| 8 | 17 | 3 | VHS | Aladdin | No | (null) | No |
| 463 | 620 | 1 | DVD Disc | Harry Potter and the Half-Blood Prince | No | (null) | Yes |
| 534 | 726 | 1 | DVD Disc | Spirited Away | No | Ryan | No |
| 571 | 807 | 1 | DVD Disc | The Original Christmas Classics | Yes | (null) | No |
Demo available at SQLFiddle
The easiest way would be to add another join to your MediaType table on FirstMediaId = MediaType.MediaId
;WITH data AS (
SELECT dbo.Release.ReleaseID
,dbo.Release.Name
,CASE WHEN Release.Compilation = 0 THEN 'No' WHEN Release.Compilation = 1 THEN 'Yes' END AS Compilation
,dbo.Release.Owner
,CASE WHEN Release.LentOut = 0 THEN 'No' WHEN Release.LentOut = 1 THEN 'Yes' END AS LentOut
,COUNT(dbo.Media.ReleaseID) AS NumberOfMedia
,MIN(dbo.Media.MediaID) AS FirstMediaID
FROM dbo.MediaType
INNER JOIN dbo.Media
ON dbo.MediaType.MediaTypeID = dbo.Media.MediaTypeID
RIGHT OUTER JOIN dbo.Release
ON dbo.Media.ReleaseID = dbo.Release.ReleaseID
GROUP BY dbo.Release.ReleaseID, dbo.Release.Name, dbo.Release.Compilation, dbo.Release.Owner, dbo.Release.LentOut
)
SELECT data.ReleaseId
,data.Name
,data.Compilation
,data.Owner
,data.LentOut
,data.NumberOfMedia
,data.FirstMediaId
,MediaType.Name as FirstMediaName
FROM data
LEFT OUTER JOIN dbo.MediaType
ON data.FirstMediaId = MediaType.MediaTypeId
for the newbie brain, this is the subquery I used
SELECT
ROW_NUMBER() OVER(PARTITION BY ReleaseID
ORDER BY MediaID) AS rn
, Media.*
FROM dbo.Media
and this is what it does (see the rn column)
| rn | MediaID | MediaTypeID | Name | ReleaseID |
|----|---------|-------------|-------------------------------------------------------------------------------------|-----------|
| 1 | 2 | 2 | Movie | 2 |
| 2 | 3 | 1 | Movie | 2 |
| 1 | 12 | 1 | Space Jam; Looney Tunes: Back in Action | 6 |
| 2 | 13 | 1 | Funky Monkey; Osmosis Jones | 6 |
| 1 | 17 | 3 | Movie | 8 |
| 1 | 620 | 1 | Movie | 463 |
| 1 | 726 | 1 | Movie | 534 |
| 1 | 807 | 1 | Rudolph the Red-Nosed Reindeer; Cricket on the Hearth | 571 |
| 2 | 808 | 1 | Frosty the Snowman; Frosty Returns | 571 |
| 3 | 809 | 1 | Santa Claus is Comin' to Town!; Mr. Magoo's Christmas Carol; The Little Drummer Boy | 571 |
| 4 | 810 | 4 | Tracks 1-7 | 571 |
Now keep only those rows with 1 in the rn column:
| rn | MediaID | MediaTypeID | Name | ReleaseID |
|----|---------|-------------|-------------------------------------------------------|-----------|
| 1 | 2 | 2 | Movie | 2 |
| 1 | 12 | 1 | Space Jam; Looney Tunes: Back in Action | 6 |
| 1 | 17 | 3 | Movie | 8 |
| 1 | 620 | 1 | Movie | 463 |
| 1 | 726 | 1 | Movie | 534 |
| 1 | 807 | 1 | Rudolph the Red-Nosed Reindeer; Cricket on the Hearth | 571 |
Then join just those rows to Releases and MediaType
Bingo
= the wanted result.
Not hard, really not hard. You really will want to learn about those window functions because they can solve heaps of problems.