I have a list of values and i want to get only the rows where the column has ONLY a combination of values:
Ex:
>CREATE TABLE User_Group(
[id] [int] NOT NULL,
[names] [varchar](255) NOT NULL,
) ON [PRIMARY]
Sample content User_Group:
1:" Joe,Jane"
2:"Jane, James,Frank"
3: "Jane, Joe,James"
I am being passed in a list of names and I want to check if the combination of names exist in the User_group table and return the rows. I want the rows ONLY if it contains the EXACT combination.
So for example, if i am given James,Jane and Joe, i want to check 2^3-1 times if James,Jane, Joe,James&Jane,James&Joe, Jane&joe, James&Jane&Joe exist in the table. And from this scenario i should only get rows 1 and 3. Row 2 is skipped because it has Frank.
I know i can do exist but not sure how to check for only that particular combination.
I am also not sure how to "loop" through all the combinations - i thought about using Java to make 2^x-1 calls with different combinations (given the scenario it is highly unlikely the combination will be >15).
I also read about "Select All" but not sure if that helps with only distinct combination either.
How can i elegantly achieve this?
This is a version that uses CTEs to do what you want.
I create a table #list that contains the lists you gave, and anoter table #search that contain each of the search terms.
declare #list table(k int, l varchar(100))
insert #list values (1,' Joe,Jane')
,(2,'Jane, James,Frank')
,(3,'Jane, Joe,James')
declare #search table(sk int,s varchar(20))
insert #search values (1,'jane'),(2,'joe'),(3,'james')
-- Top level CTE to remove spaces, and each term is surrounded by its own commas.
;with cte as (
select k,','+replace(replace(l,',',',,'),' ','')+',' l from #list
)
-- Go through the search terms recursively and remove the search terms, with their surrounding commas
,cte2 as(
select cte.k, replace(cte.l,','+s+',','') as l, s.sk, s.s
from cte
join #search s on sk=1
union all
select cte2.k, replace(cte2.l,','+s.s+',','') as l, s.sk ,s.s
from cte2
join #search s on s.sk=cte2.sk+1
)
-- Find any that result in zero length
select distinct k from cte2 where len(l)=0
Related
I have a Redshift table with the following column
How can I extract the value starting by cat_ from this column please (there is only one for each row and at different position in the array)?
I want to get those results:
cat_incident
cat_feature_missing
cat_duplicated_request
Thanks!
There is no easy way to extract multiple values from within one column in SQL (or at least not in the SQL used by Redshift).
You could write a User-Defined Function (UDF) that returns a string containing those values, separated by newlines. Whether this is acceptable depends on what you wish to do with the output (eg JOIN against it).
Another option is to pre-process the data before it is loaded into Redshift, to put this information in a separate one-to-many table, with each value in its own row. It would then be trivial to return this information.
You can do this using tally table (table with numbers). Check this link on information how to create this table: http://www.sqlservercentral.com/articles/T-SQL/62867/
Here is example how you would use it. In real life you should replace temporary #tally table with a permanent one.
--create sample table with data
create table #a (tags varchar(500));
insert into #a
select 'blah,cat_incident,mcr_close_ticket'
union
select 'blah-blah,cat_feature_missing,cat_duplicated_request';
--create tally table
create table #tally(n int);
insert into #tally
select 1
union select 2
union select 3
union select 4
union select 5
;
--get tags
select * from
(
select TRIM(SPLIT_PART(a.tags, ',', t.n)) AS single_tag
from #tally t
inner join #a a ON t.n <= REGEXP_COUNT(a.tags, ',') + 1 and n<1000
)
where single_tag like 'cat%'
;
Thanks!
In the end I managed to do it with the following query:
SELECT SUBSTRING(SUBSTRING(tags, charindex('cat_', tags), len(tags)), 0, charindex(',', SUBSTRING(tags, charindex('cat_', tags), len(tags)))) tags
FROM table
Have two tables, table 1 with columns W_ID and word. Table 2 with column N_ID and note. Have to list all the NID where words found in table 1 word column contains in Note column (easy part) and also list those words in another column without duplicating the N_ID. Which means using STUFF to concatenate all the words found in Note column for that particular N_ID. I tried using
FULL TEXT INDEX using CONTAIN
But it only allows to search for one word at a time. Any suggestions how I can use a while loop to achieve this.
If there is a maximum number of words you want displayed for N_ID, you can pivot this. You could have them in a single column by concatenating them, but I would recommend against that. Here is a pivot that supports up to 4 words per N_ID. You can adjust it as needed. You can view the SQL Fiddle for this here.
SELECT
n_id,
[1] AS word_1,
[2] AS word_2,
[3] AS word_3,
[4] AS word_4
FROM (
SELECT
n_id,
word,
ROW_NUMBER() OVER (PARTITION BY n_id ORDER BY word) AS rn
FROM tbl2
JOIN tbl1 ON
tbl2.note LIKE '%'+tbl1.word+'[ ,.?!]%'
) AS source_table
PIVOT (
MAX(word)
FOR rn IN ([1],[2],[3],[4])
) AS pivot_table
*updated the join to prevent look for a space or punctuation to declare the end of a word.
You can join your tables together based on a postive result from the charindex function.
In SQL 2017 you can run:
SELECT n_id, string_agg(word)
FROM words
inner join notes on 0 < charindex(words.word, notes.note);
Prior to SQL 2017, there is no string_agg so you'll need to use stuff, which is trickier:
select
stuff((
SELECT ', ' + word
FROM words
where 0 < charindex(words.word, notes.note)
FOR XML PATH('')
), 1, 2, '')
from notes;
I used the following schema:
CREATE table WORDS
(W_ID int identity primary key
,word varchar(100)
);
CREATE table notes
(N_ID int identity primary key
,note varchar(1000)
);
insert into words (word) values
('No'),('Nope'),('Nah');
insert into notes (note) values
('I am not going to do this. Nah!!!')
,('It is OK.');
I have some question in Sqlserver2012. I have a table that contains a filed that save who System Used from this information and separated by ',', I want to set into parameter the name of Systems and query the related rows:
declare #System nvarchar(50)
set #System ='BPM,SEM'
SELECT *
FROM dbo.tblMeasureCatalog t1
where ( ( select Upper(value) from dbo.split(t1.System,','))
= any( select Upper(value) from dbo.split(#System,',')))
dbo.split is a function to return systems in separated rows
Forgetting for a second that storing delimited lists in a relational database is abhorrent, you can do it using a combination of INTERSECT and EXISTS, for example:
DECLARE #System NVARCHAR(50) = 'BPM,SEM';
DECLARE #tblMeasureCatalog TABLE (System VARCHAR(MAX));
INSERT #tblMeasureCatalog VALUES ('BPM,XXX'), ('BPM,SEM'), ('XXX,SEM'), ('XXX,YYY');
SELECT mc.System
FROM #tblMeasureCatalog AS mc
WHERE EXISTS
( SELECT Value
FROM dbo.Split(mc.System, ',')
INTERSECT
SELECT Value
FROM dbo.Split(#System, ',')
);
Returns
System
---------
BPM,XXX
BPM,SEM
XXX,SEM
EDIT
Based on your question stating "Any" I assumed that you wanted rows where the terms matched any of those provided, based on your comment I now assume you want records where the terms match all. This is a fairly similar approach but you need to use NOT EXISTS and EXCEPT instead:
Now all is still quite ambiguous, for example if you search for "BMP,SEM" should it return a record that is "BPM,SEM,YYY", it does contain all of the searched terms, but it does contain additional terms too. So the approach you need depends on your requirements:
DECLARE #System NVARCHAR(50) = 'BPM,SEM,XXX';
DECLARE #tblMeasureCatalog TABLE (System VARCHAR(MAX));
INSERT #tblMeasureCatalog
VALUES
('BPM,XXX'), ('BPM,SEM'), ('XXX,SEM'), ('XXX,YYY'),
('SEM,BPM'), ('SEM,BPM,XXX'), ('SEM,BPM,XXX,YYY');
-- METHOD 1 - CONTAINS ALL SEARCHED TERMS BUT CAN CONTAIN ADDITIONAL TERMS
SELECT mc.System
FROM #tblMeasureCatalog AS mc
WHERE NOT EXISTS
(
SELECT Value
FROM dbo.Split(#System, ',')
EXCEPT
SELECT Value
FROM dbo.Split(mc.System, ',')
);
-- METHOD 2 - ONLY CONTAINS ITEMS WITHIN THE SEARCHED TERMS, BUT NOT
-- NECESSARILY ALL OF THEM
SELECT mc.System
FROM #tblMeasureCatalog AS mc
WHERE NOT EXISTS
( SELECT Value
FROM dbo.Split(mc.System, ',')
EXCEPT
SELECT Value
FROM dbo.Split(#System, ',')
);
-- METHOD 3 - CONTAINS ALL ITEMS IN THE SEARCHED TERMS, AND NO ADDITIONAL ITEMS
SELECT mc.System
FROM #tblMeasureCatalog AS mc
WHERE NOT EXISTS
( SELECT Value
FROM dbo.Split(#System, ',')
EXCEPT
SELECT Value
FROM dbo.Split(mc.System, ',')
)
AND LEN(mc.System) = LEN(#System);
You have a problem with your data structure because you are storing lists of things in a comma-delimited list. SQL has a great data structure for storing lists. It goes by the name "table". You should have a junction table with one row per "measure catalog" and "system".
Sometimes, you are stuck with other people's really bad design decisions. One solution is to use split(). Here is one method:
select mc.*
from dbo.tblMeasureCatalog mc
where exists (select 1
from dbo.split(t1.System, ',') t1s join
dbo.split(#System, ',') ss
on upper(t1s.value) = upper(ss.value)
);
you can try this :
declare #System nvarchar(50)
set #System ='BPM,SEM'
SELECT * from dbo.tblMeasureCatalog t1 inner join dbo.Split (#System ,',') B on t1.it=B.items
I'm a longtime SAS programmer and we're looking at moving our system from SAS to another platform. I've only a very basic knowledge of SQL the marketing folks talk about using SQL a lot but I wonder how it might do some things we need done. For instance, we have files with up to 50 million rows of vaccination records for each vaccine that was administered to a patient. Some vaccines are actually a combination vaccine that represent 2-4 different types of vaccines. The type of vaccine is based on the value of CVX. Using a do-loop it's fairly simple to do this in SAS, but I've no idea of how it might be done in SQL. It's be safe to assume that we have all the CVX codes in a table with 1 to 4 vaccine types that need to be generated. But how would you do it in SQL?
Thanks,
Steve
I don't know anything about your schema, but it sounds like you're looking to split multiple columns into rows. You can use either CROSS APPLY or the UNPIVOT operator. Here is an example that takes a contrived test table and splits it into separate rows per key:
create table #Test
(
Test_Key int identity(1,1) primary key clustered,
Test_A int,
Test_B int,
Test_C int
)
declare #n int
set #n = 1
while #n < 10000
begin
insert into #Test (Test_A, Test_B, Test_C)
select #n * 5 + 1, #n * 5 + 2, #n * 5 + 3
set #n = #n + 1
end
select * from #Test
-- this example converts the columns into rows using CROSS APPLY
-- this may be slightly less expensive than the UNPIVOT example below
select
F_Key,
F_Value
from #Test
cross apply
(
values
(Test_Key, Test_A), -- 1st row is Test_A
(Test_Key, Test_B), -- 2nd row is Test_B
(Test_Key, Test_C) -- 3rd row is Test_C
) as F(F_Key, F_Value)
-- this example converts the columns into rows using the UNPIVOT operator
select
Test_Key, TestKey
from #Test
unpivot
(TestKey for Test_Type in (Test_A, Test_B, Test_C)) as C
drop table #Test
In the code_list CTE in this query I have a row constructor that will eventually take any number of arguments. The column icd in the patient_codes CTE is a five digit identifier that is most descriptive that the three digit codes that the row constructor has. The table icd_patient has a 100 million rows so for performance's sake, I would like to filer the rows on this table before I do any further work. I have
;with code_list(code_list)
as
(
select x.code_list
from (values ('70700'),('25002')) as x(code_list)
),patient_codes
as
(
select distinct icd,pat_id,id
from icd_patient
where icd in (select icd from code_list)
)
select distinct pat_id from patient_codes
The problem is, however, is that in the icd_patient table all of the icd columns are five digit and more descriptive. If I look at the execution plan of this query it's pretty streamlined. If I do
;with code_list(code_list)
as
(
select x.code_list
from (values ('70700'),('25002')) as x(code_list)
),patient_codes
as
(
select substring(icd,1,3) as icd,pat_id
from icd_patient2
where substring(icd,1,3) in (select * from code_list)
)
select * from patient_codes
this if course has a large performance impact because of the substring expression in the where clause. Does something akin to like in exist so I can take advantage of my indexes?
Index on icd_patient
CREATE NONCLUSTERED INDEX [ix_icd_patient] ON [dbo].[icd_patient2]
(
[pat_id] ASC
)
INCLUDE ( [id],
This much simpler query should be better than (or, at worst, the same as) your existing query.
select pat_id
FROM dbo.icd_patient
where icd LIKE '707%'
OR icd LIKE '250%'
GROUP BY pat_id;
Note that sargability only matters if there is actually an index on this column.
An alternative (since OR can sometimes give the optimizer fits):
SELECT pat_id FROM
(
SELECT pat_id
FROM dbo.icd_patient
WHERE icd LIKE '707%'
UNION ALL
SELECT pat_id
FROM dbo.icd_patient
WHERE icd LIKE '250%'
) AS x
GROUP BY pat_id;
To make this extensible beyond a handful of OR conditions, I would use a table-valued parameter (TVP).
CREATE TYPE dbo.StringPatterns AS TABLE(s VARCHAR(3) PRIMARY KEY);
Then your stored procedure could say:
CREATE PROCEDURE dbo.whatever
#sp dbo.StringPatterns READONLY
AS
BEGIN
SET NOCOUNT ON;
SELECT p.pat_id
FROM dbo.icd_patient AS p
INNER JOIN #sp AS sp
ON p.pat_id LIKE sp.s + '%'
GROUP BY p.pat_id;
END
Then you can pass in your set of three-character substrings from a DataTable or other collection in C#. From T-SQL just as an example:
DECLARE #p dbo.StringPatterns;
INSERT #p VALUES('707'),('250');
EXEC dbo.whatever #sp = #p;
Something like like in does not exist. The following is sargable:
select *
from icd_patient
where icd like '70700%' or
icd like '25002%'
Because like with a constant initial substring is a special case for SQL Server. This does not work when the strings on the right are variables.
One solution is to create an indexed view on the icd_patient table with an index on the first five characters of the icd code.
Using "IN" makes that part of a command non-sargable on both sides. End of discussion.
Saying he fixes it using substring, completely changes what it would return while it remains non sarged.
Any "fix" should exactly match results. The actual fix is to join the cte so the five characters match or put three characters in the cte and match that in a join or put 4 characters in the cte where the fourth is "%" and join matching by using LIKE
Using a "like" that starts with "%" increases the complexity of the search, but it would still use the index to find the value because parsing the index should use less reading by only getting the full table row when a search is successful.