SQL Delete Where Not In with Composite Key - sql

I have a system integration project which needs to CRUD from one DB to another. Not especially complicated. However, when it comes to deleting rows which exist in the target but not in the source, I ran into a little trouble. The standard patterns include: LEFT JOIN, NOT EXISTS or NOT IN. I chose the LEFT JOIN. My 'Phone' table uses a composite key, Employee 'Id' and the PhoneType: Work, Home, Mobile, etc. The standard left join will delete ANY target Phone number NOT in the source. This clears out the whole table. NOTE: I am updating only records which have changed since the last update, NOT the whole target & source. So, I wrote a fix which I suspect is really poor SQL:
-- SOURCE
DECLARE #tmpPhones TABLE(Id varchar(8), PhoneType int, PhoneNumber varchar(30), PRIMARY KEY (Id, PhoneType))
INSERT into #tmpPhones values
('TEST123', 1, '12345678'),
('TEST123', 2, '12345678'),
('TEST123', 3, '12345678')
-- TARGET
DECLARE#Phone TABLE( Id varchar(8), PhoneType int, PhoneNumber varchar(30), PRIMARY KEY (Id, PhoneType))
INSERT into #Phone values
('TEST123', 1, '12345678'), <-- Exists in both, leave
('TEST123', 2, '12345678'), <-- Exists in both, leave
('TEST123', 3, '12345678'), <-- Exists in both, leave
('TEST123', 4, '12345678'), <-- ONLY delete this one!
('TEST456', 2, '12345678'), <-- Ignore this employee Id
('TEST456', 3, '12345678'), ""
('TEST456', 4, '12345678') ""
DELETE p
FROM #Phone p
LEFT JOIN #tmpPhones t
ON t.Id = p.Id AND t.PhoneType = p.PhoneType
WHERE t.Id IS NULL AND t.PhoneType IS NULL
AND p.Id IN (SELECT Id FROM #tmpPhones) <-- a sad hack?
This works, but I feel like there is a better way to make sure we are only deleting records for THIS employee, not all the others.
Any suggestions?

Use exists.
DELETE p
FROM #Phone p
where exists (select 1 from #tmpPhones where Id = p.Id)
AND not exists (select 1 from #tmpPhones where PhoneType = p.PhoneType)
Edit: Deleting using cte.
with todelete as (
select id,phonetype from phone
except
select id,phonetype from tmpphones t
where exists (select 1 from phone where id = t.id)
)
delete from phone
where exists (select 1 from todelete where phone.id = id and phone.phonetype = phonetype)

I think two exists statements pretty much capture the logic: as you describe it
DELETE p
FROM #Phone p
WHERE EXISTS (SELECT 1 FROM #tmpPhone t WHERE t.id = p.id) AND
NOT EXISTS (SELECT 1 FROM #tmpPhone t WHERE t.id = p.id AND t.PhoneType = p.PhoneType) ;

merge seems to work ok - but you still need to see if the id is in your reference set, I don't see a clean way around that
MERGE #Phone AS TGT
USING (
SELECT * FROM #tmpPhones
) AS SRC
ON TGT.ID=SRC.ID AND TGT.PHONETYPE=SRC.PHONETYPE
WHEN NOT MATCHED BY SOURCE AND tgt.id IN (SELECT id FROM #tmpPhones) THEN DELETE;

Related

SQL How to get full Array in a temporary table with condition?

Well, I am really sorry because my explanation was so poor. Thank you for all the answers.
I will explain better what should be the output and what is my question.
So, first of I have an array of tagCodes like ('code0','code1','code2').
Then I have a table that contains Codes and TagTypeId.
I would like to get into a temporary table all the codes I passed in the array with their TagTypeId. So a table like:
Code
TagTypeId
903420012408181609019A18
2456
903420012408181609019A18
2135
TestCodeNull
null
So my attempt was this one:
SELECT Tags.Code AS tagCode, Tags.TagTypeId, TagTypes.Code AS tagType
INTO #TempTable
FROM Tags JOIN TagTypes ON Tags.TagTypeId = TagTypes.Id
WHERE Tags.Code IN ('903420012408181609019A18','90341808151313061101E938', 'TestCodeNull')
SELECT * FROM #TempTable;
But I dont get the codes that are not in the Tag table.
I did this an it seems to be working as intended:
CREATE TABLE #TestTable (tagCode NVARCHAR(25) NOT NULL, TagTypeId INT NULL, tagType NVARCHAR(MAX))
INSERT INTO #TestTable (tagCode) VALUES ('903420012408181609019A18'),('00007E08190D0A34E1F524D0'),('00007E08190D0B25E1F5A98B')
UPDATE #TestTable SET TagTypeId = Tags.TagTypeId, tagType = TagTypes.Code FROM #TestTable
LEFT JOIN Tags ON (#TestTable.tagCode = Tags.Code)
LEFT JOIN TagTypes ON (Tags.TagTypeId = TagTypes.Id)
SELECT * FROM #TestTable;
I think what you mean that 'TestCodeNull' does not exist in tags so you want to show null for 'TestCodeNull' in which case a join may be more appropriate. for example
SELECT S.CODE,Tags.Code AS tagCode, Tags.TagTypeId,
TagTypes.Code AS tagType
INTO #TempTable
FROM (select '903420012408181609019A18' code
union all select '90341808151313061101E938'
union all select 'TestCodeNull') s
left join Tags on tags.code = s.code
left JOIN TagTypes ON Tags.TagTypeId = TagTypes.Id
SELECT * FROM #TempTable;

SQL query based on values from another tables

I need to put external courses to "ext" folder.
Which course is external? -External course is course whose vendor doesn't have a client_id.
So first we check if vendor for our course has a client_id, we will check for that in vendor table. If our vendor doesn't have client_id (client_id = NULL) then we want to grab an ID of "ext" record from folder table and insert it to folder_id column. If our vendor has a client_id we dont do anything, so second row in course table under folder_id will be empty. Folder_id column should be populated like this: 1.row = 2, 2.row = (empty), 3.row = 2, 4.row = 2, 5.row = 2.
Here is the code that I have for now but I'm getting 'column reference "id" is ambiguous' at line 2. I dont know if code is good, I'm not that good with sql.
(prefix 't' in code means table)
UPDATE t_course SET folder_id = (SELECT id FROM t_folder WHERE code IS NULL)
WHERE id IN (SELECT id
FROM t_course course
JOIN t_vendor vendor ON vendor.id = course.vendor_id
WHERE vendor.client_id IS NULL);
I suspect you really want EXISTS:
UPDATE t_course c
SET folder_id = (SELECT f.id FROM t_folder f WHERE f.code IS NULL)
WHERE EXISTS (SELECT 1
FROM t_vendor vendor
WHERE v.id = c.vendor_id AND v.client_id IS NULL
);

remove Union from query

I have query to get category and subcategories of same category. I used below query:
Select Id from category where name like '%events%' and deleted = 0 and published = 1
UNION
SELECT Id from category where parentcategoryid = (Select id from category where name like '%events%' and deleted = 0 and published = 1)
I do not want to use UNION, want to use Join only. But not getting how i can achieve. Below is the table structure. Please help me. thanks in Advance
The following might solve your problem:
DECLARE #t TABLE (id int, name varchar(20), parentcatid varchar(200))
insert into #t values(23,'Christmas', 34)
,(29,'Birthday', 34)
,(31,'New year', 34)
,(34,'Events', 0)
,(35,'gfrhrt', 0)
;WITH cte AS(
Select Id from #t
where name like '%events%' --and deleted = 0 and published = 1
),
cte2 AS (
SELECT a.id AS tId, cpa.id, cId2.id AS idPa
from #t a
LEFT JOIN cte AS cId ON cId.Id = a.Id
FULL OUTER JOIN #t AS cPa ON cPa.parentcatid = cId.Id
LEFT JOIN cte cId2 ON cId2.id = cPa.id
WHERE cId.Id IS NOT NULL OR cPa.Id IS NOT NULL
)
SELECT id
FROM cte2
WHERE tId IS NOT NULL
OR id = idPa
The idea is to get all required IDs within the cte and then get all categories, where either ID oder ParentID match the IDs from the cte. However, depending on the size of your table, you might want to add further filters to the cte.

How to UPDATE pivoted table in SQL SERVER

I have flat table which I have to join using EAN attribute with my main table and update gid (id of my main table).
id attrib value gid
1 weight 10 NULL
1 ean 123123123112 NULL
1 color blue NULL
2 weight 5 NULL
2 ean 331231313123 NULL
I was trying to pivot ean rows into column, next join on ean both tables, and for this moment everything works great.
--update SideTable
--set gid = ab_id
select gid, ab_id
from SideTable
pivot (max (value) for attrib in ([EAN],[MPN])) as b
join MainTable as c
on c.ab_ean = b.EAN
where b.EAN !='' AND c.ab_archive = '0'
When I am selecting both id columns is okey, but when I am uncomment first lines and delete select whole table is set with first gid from my main table.
It have to set my main id into all attributes where ID where ean is matched from my main table.
I am sorry for my terrible english but I hope someone can help me, with that.
The reason your update does not work is that you don't have any link between your source and target for the update, although you reference sidetable in the FROM clause, this is effectively destroyed by the PIVOT function, leaving no link back to the instance of SideTable that you are updating. Since there is no link, all rows are updated with the same value, this will be the last value encountered in the FROM.
This can be demonstrated by running the following:
DECLARE #S TABLE (ID INT, Attrib VARCHAR(50), Value VARCHAR(50), gid INT);
INSERT #S
VALUES
(1, 'weight', '10', NULL), (1, 'ean', '123123123112', NULL), (1, 'color', 'blue', NULL),
(2, 'weight', '5', NULL), (2, 'ean', '331231313123', NULL);
SELECT s.*
FROM #S AS s
PIVOT (MAX(Value) FOR attrib IN ([EAN],[MPN])) AS pvt;
You clearly have a table aliased s in the FROM clause, however because you have used pivot you cannot use SELECT s*, you get the following error:
The column prefix 's' does not match with a table name or alias name used in the query.
You haven't provided sample data for your main table, but I am about 95% certain your PIVOT is not needed, I think you can get your update using just normal JOINs:
UPDATE s
SET gid = ab_id
FROM SideTable AS s
INNER JOIN SideTable AS ean
ON ean.ID = s.ID
AND ean.attrib = 'ean'
INNER JOIN MainTable AS m
ON m.ab_EAN = ean.Value
WHERE m.ab_archive = '0'
AND m.ab_EAN != '';
As per comment to the question, you need to use update + select statement.
A standard version looks like:
UPDATE
T
SET
T.col1 = OT.col1,
T.col2 = OT.col2
FROM
Some_Table T
INNER JOIN
Other_Table OT
ON
T.id = OT.id
WHERE
T.col3 = 'cool'
As to your needs:
update a
set a.gid = p.ab_id
from SideTable As a
Inner join (
select gid, ab_id
from SideTable
pivot (max (value) for attrib in ([EAN],[MPN])) as b
join MainTable as c
on c.ab_ean = b.EAN
where b.EAN !='' AND c.ab_archive = '0') p ON a.ean = p.EAN
try and break it down a bit more like this..
update SideTable
set SideTable.gid = p.ab_id
FROM
(
select gid, ab_id
from SideTable
pivot (max (value) for attrib in ([EAN],[MPN])) as b
join MainTable as c
on c.ab_ean = b.EAN
where b.EAN !='' AND c.ab_archive = '0'
) p
WHERE p.EAN = SideTable.EAN

querying 2 tables with the same spec for the differences

I recently had to solve this problem and find I've needed this info many times in the past so I thought I would post it. Assuming the following table def, how would you write a query to find all differences between the two?
table def:
CREATE TABLE feed_tbl
(
code varchar(15),
name varchar(40),
status char(1),
update char(1)
CONSTRAINT feed_tbl_PK PRIMARY KEY (code)
CREATE TABLE data_tbl
(
code varchar(15),
name varchar(40),
status char(1),
update char(1)
CONSTRAINT data_tbl_PK PRIMARY KEY (code)
Here is my solution, as a view using three queries joined by unions. The diff_type specified is how the record needs updated: deleted from _data(2), updated in _data(1), or added to _data(0)
CREATE VIEW delta_vw AS (
SELECT feed_tbl.code, feed_tbl.name, feed_tbl.status, feed_tbl.update, 0 as diff_type
FROM feed_tbl LEFT OUTER JOIN
data_tbl ON feed_tbl.code = data_tbl.code
WHERE (data_tbl.code IS NULL)
UNION
SELECT feed_tbl.code, feed_tbl.name, feed_tbl.status, feed_tbl.update, 1 as diff_type
FROM data_tbl RIGHT OUTER JOIN
feed_tbl ON data_tbl.code = feed_tbl.code
where (feed_tbl.name <> data_tbl.name) OR
(data_tbl.status <> feed_tbl.status) OR
(data_tbl.update <> feed_tbl.update)
UNION
SELECT data_tbl.code, data_tbl.name, data_tbl.status, data_tbl.update, 2 as diff_type
FROM feed_tbl LEFT OUTER JOIN
data_tbl ON data_tbl.code = feed_tbl.code
WHERE (feed_tbl.code IS NULL)
)
UNION will remove duplicates, so just UNION the two together, then search for anything with more than one entry. Given "code" as a primary key, you can say:
edit 0: modified to include differences in the PK field itself
edit 1: if you use this in real life, be sure to list the actual column names. Dont use dot-star, since the UNION operation requires result sets to have exactly matching columns. This example would break if you added / removed a column from one of the tables.
select dt.*
from
data_tbl dt
,(
select code
from
(
select * from feed_tbl
union
select * from data_tbl
)
group by code
having count(*) > 1
) diffs --"diffs" will return all differences *except* those in the primary key itself
where diffs.code = dt.code
union --plus the ones that are only in feed, but not in data
select * from feed_tbl ft where not exists(select code from data_tbl dt where dt.code = ft.code)
union --plus the ones that are only in data, but not in feed
select * from data_tbl dt where not exists(select code from feed_tbl ft where ft.code = dt.code)
I would use a minor variation in the second union:
where (ISNULL(feed_tbl.name, 'NONAME') <> ISNULL(data_tbl.name, 'NONAME')) OR
(ISNULL(data_tbl.status, 'NOSTATUS') <> ISNULL(feed_tbl.status, 'NOSTATUS')) OR
(ISNULL(data_tbl.update, '12/31/2039') <> ISNULL(feed_tbl.update, '12/31/2039'))
For reasons I have never understood, NULL does not equal NULL (at least in SQL Server).
You could also use a FULL OUTER JOIN and a CASE ... END statement on the diff_type column along with the aforementioned where clause in querying 2 tables with the same spec for the differences
That would probably achieve the same results, but in one query.