How To Extract specific key Value from the string that has key value pairs

Question

Post reply

How To Extract specific key Value from the string that has key value pairs

Zond Sita

SSC Veteran

Points: 275

January 24, 2024 at 11:53 pm

Go to Answer

#4367290

How To Extract Values from the string that is key value pairs with different formats

I have a table called DataForce. with column DataVal (Nvarchar 2000).

it has below 3 records

How to get CID Value, ID Value, Email Value, Zip from below strings please advise?

{'ani':'80052242','dnis':'90425935','cid':'4601923812','Id':'046067ce-7fec-4af7-bb51-e4d31b4720','No':'3048','FN':'Bod Conner','Ent':'2012-08-27','Email':'sounage@nefcom.net','Add':'P.o. Box 146','Zip':'31040','Not':'English','Off':'FL','ACode':'','OffId':'0002 Home Office FL','Except':''}, 
{"ani":"30926539","dnis":"33300180","cid":"4594296356","Id":"a2a5bac2-b6ad-4239-95b2-2da730c2b6","No":"93872","FN":"Martinael u0026 Kar Schmale","Ent":"2023-07-08","Email":"martinhael.d.schmale@gmail.com","Add":"552 WPlainview cir","Zip":"13498","Not":"","Off":"FL","ACode":"1|a2a5bac2-b6ad-4239-95b2-2da780c2b6|0061FL|V|000000000000000000|00000000","OffId":"0061 Rand FL","Except":""},
{ani:63179364,dnis:88825454,cid:4563900520,Id:c1144de0-1945-4862-b057-fe25eabe73,No:97158,FN:Bombay and Marsha Osrow,Ent:2023-11-03,Email:bombayrow7634@aol.com,Add:10014 SW Vitalia Court,Zip:14987,Not:English,Off:FL,ACode:,OffId:0021 Zanfini FL,Except:},

Thank you in advance

asita

Phil Parkin

SSC Guru

Points: 246960

January 25, 2024 at 9:56 am

Answer

#4367524

Here is the STRING_SPLIT version

SELECT
    jn.Id
,   cid = MAX (REPLACE (IIF(c2.ordinal = 3, c2.value, ''), 'cid:', ''))
,   Id = MAX (REPLACE (IIF(c2.ordinal = 4, c2.value, ''), 'Id:', ''))
,   Email = MAX (REPLACE (IIF(c2.ordinal = 8, c2.value, ''), 'Email:', ''))
FROM #JSONish jn
    CROSS APPLY
(SELECT Cleaned = REPLACE (REPLACE (jn.Dataval, '''', ''), '"', '')) c1
    CROSS APPLY
(SELECT ss.* FROM STRING_SPLIT(c1.Cleaned, ',', 1) ss) c2
GROUP BY jn.Id
ORDER BY jn.Id;

The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin

January 29, 2024 at 11:03 am

This was removed by the editor as SPAM

Viewing 15 posts - 1 through 15 (of 17 total)

You must be logged in to reply to this topic. Login to reply

Zond Sita SSC Veteran Points: 275 More actions · Answer 1

Tried this below code but no luck 🙁

Can somebody please help. Thank you in advance

SELECT 
Dataval,
    CASE 
        WHEN LEN(SUBSTRING(SUBSTRING(Dataval, CHARINDEX('"Id":"', Dataval) + 6, LEN(Dataval)), 1, CASE WHEN CHARINDEX('","No"', SUBSTRING(Dataval, CHARINDEX('"Id":"', Dataval) + 6, LEN(Dataval))) > 0 THEN CHARINDEX('","No"', SUBSTRING(Dataval, CHARINDEX('"Id":"', Dataval) + 6, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)) > 70
            THEN SUBSTRING(SUBSTRING(Dataval, CHARINDEX('"Id":"', Dataval) + 6, LEN(Dataval)), 1, CASE WHEN CHARINDEX('","No"', SUBSTRING(Dataval, CHARINDEX('"Id":"', Dataval) + 6, LEN(Dataval))) > 0 THEN CHARINDEX('","No"', SUBSTRING(Dataval, CHARINDEX('"Id":"', Dataval) + 6, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)
        ELSE SUBSTRING(SUBSTRING(Dataval, CHARINDEX(',Id:', Dataval) + 5, LEN(Dataval)), 1, CASE WHEN CHARINDEX(',No', SUBSTRING(Dataval, CHARINDEX(',Id:', Dataval) + 5, LEN(Dataval))) > 0 THEN CHARINDEX(',No:', SUBSTRING(Dataval, CHARINDEX(',Id:', Dataval) + 5, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)
    END AS COSTID,
    CASE 
        WHEN LEN(SUBSTRING(SUBSTRING(Dataval, CHARINDEX('"OffId":"', Dataval) + 9, LEN(Dataval)), 1, CASE WHEN CHARINDEX('","Except"', SUBSTRING(Dataval, CHARINDEX('"OffId":"', Dataval) + 9, LEN(Dataval))) > 0 THEN CHARINDEX('","Except"', SUBSTRING(Dataval, CHARINDEX('"OffId":"', Dataval) + 9, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)) > 70
            THEN SUBSTRING(SUBSTRING(Dataval, CHARINDEX('"OffId":"', Dataval) + 9, LEN(Dataval)), 1, CASE WHEN CHARINDEX('","Except"', SUBSTRING(Dataval, CHARINDEX('"OffId":"', Dataval) + 9, LEN(Dataval))) > 0 THEN CHARINDEX('","Except"', SUBSTRING(Dataval, CHARINDEX('"OffId":"', Dataval) + 9, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)
        ELSE SUBSTRING(SUBSTRING(Dataval, CHARINDEX(',OffId:', Dataval) + 7, LEN(Dataval)), 1, CASE WHEN CHARINDEX(',Except', SUBSTRING(Dataval, CHARINDEX(',OffId:', Dataval) + 7, LEN(Dataval))) > 0 THEN CHARINDEX(',Except', SUBSTRING(Dataval, CHARINDEX(',OffId:', Dataval) + 7, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)
    END AS Branch,
    CASE 
        WHEN LEN(SUBSTRING(SUBSTRING(Dataval, CHARINDEX('"Email":"', Dataval) + 8, LEN(Dataval)), 1, CASE WHEN CHARINDEX('","Add"', SUBSTRING(Dataval, CHARINDEX('"Email":"', Dataval) + 8, LEN(Dataval))) > 0 THEN CHARINDEX('","Add"', SUBSTRING(Dataval, CHARINDEX('"Email":"', Dataval) + 8, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)) > 70
            THEN SUBSTRING(SUBSTRING(Dataval, CHARINDEX('"Email":"', Dataval) + 8, LEN(Dataval)), 1, CASE WHEN CHARINDEX('","Add"', SUBSTRING(Dataval, CHARINDEX('"Email":"', Dataval) + 8, LEN(Dataval))) > 0 THEN CHARINDEX('","Add"', SUBSTRING(Dataval, CHARINDEX('"Email":"', Dataval) + 8, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)
        ELSE SUBSTRING(SUBSTRING(Dataval, CHARINDEX(',Email:', Dataval) + 8, LEN(Dataval)), 1, CASE WHEN CHARINDEX(',Add:', SUBSTRING(Dataval, CHARINDEX(',Email:', Dataval) + 8, LEN(Dataval))) > 0 THEN CHARINDEX(',Add:', SUBSTRING(Dataval, CHARINDEX(',Email:', Dataval) + 8, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)
    END AS ContactEmail,
    CASE 
        WHEN LEN(SUBSTRING(SUBSTRING(Dataval, CHARINDEX('"Zip":"', Dataval) + 6, LEN(Dataval)), 1, CASE WHEN CHARINDEX('","Not"', SUBSTRING(Dataval, CHARINDEX('"Zip":"', Dataval) + 6, LEN(Dataval))) > 0 THEN CHARINDEX('","Not"', SUBSTRING(Dataval, CHARINDEX('"Zip":"', Dataval) + 6, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)) > 70
            THEN SUBSTRING(SUBSTRING(Dataval, CHARINDEX('"Zip":"', Dataval) + 6, LEN(Dataval)), 1, CASE WHEN CHARINDEX('","Not"', SUBSTRING(Dataval, CHARINDEX('"Zip":"', Dataval) + 6, LEN(Dataval))) > 0 THEN CHARINDEX('","Not"', SUBSTRING(Dataval, CHARINDEX('"Zip":"', Dataval) + 6, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)
        ELSE SUBSTRING(SUBSTRING(Dataval, CHARINDEX(',Zip:', Dataval) + 5, LEN(Dataval)), 1, CASE WHEN CHARINDEX(',Not', SUBSTRING(Dataval, CHARINDEX(',Zip:', Dataval) + 5, LEN(Dataval))) > 0 THEN CHARINDEX(',Not', SUBSTRING(Dataval, CHARINDEX(',Zip:', Dataval) + 5, LEN(Dataval))) - 1 ELSE LEN(Dataval) END)
    END AS PostalCode
FROM DataForce c
;

Phil Parkin SSC Guru Points: 246960 More actions · Answer 2

I've done the first few for you. You should be able to add the others using similar logic.

DROP TABLE IF EXISTS #JSONish;

CREATE TABLE #JSONish
(
    Id INT IDENTITY(1, 1)
   ,Dataval NVARCHAR(MAX) NOT NULL
);

INSERT #JSONish
(
    Dataval
)
VALUES
(N'{''ani'':''80052242'',''dnis'':''90425935'',''cid'':''4601923812'',''Id'':''046067ce-7fec-4af7-bb51-e4d31b4720'',''No'':''3048'',''FN'':''Bod Conner'',''Ent'':''2012-08-27'',''Email'':''sounage@nefcom.net'',''Add'':''P.o. Box 146'',''Zip'':''31040'',''Not'':''English'',''Off'':''FL'',''ACode'':'''',''OffId'':''0002 Home Office FL'',''Except'':''''}, ' -- Dataval - nvarchar(max)
    )
,(N'{"ani":"30926539","dnis":"33300180","cid":"4594296356","Id":"a2a5bac2-b6ad-4239-95b2-2da730c2b6","No":"93872","FN":"Martinael u0026 Kar Schmale","Ent":"2023-07-08","Email":"martinhael.d.schmale@gmail.com","Add":"552 WPlainview cir","Zip":"13498","Not":"","Off":"FL","ACode":"1|a2a5bac2-b6ad-4239-95b2-2da780c2b6|0061FL|V|000000000000000000|00000000","OffId":"0061 Rand FL","Except":""},')
,(N'{ani:63179364,dnis:88825454,cid:4563900520,Id:c1144de0-1945-4862-b057-fe25eabe73,No:97158,FN:Bombay and Marsha Osrow,Ent:2023-11-03,Email:bombayrow7634@aol.com,Add:10014 SW Vitalia Court,Zip:14987,Not:English,Off:FL,ACode:,OffId:0021 Zanfini FL,Except:},');

SELECT jn.Id
      ,cid = MAX (REPLACE (IIF(c2.ItemNumber = 3, c2.Item, ''), 'cid:', ''))
      ,Id = MAX (REPLACE (IIF(c2.ItemNumber = 4, c2.Item, ''), 'Id:', ''))
      ,Email = MAX (REPLACE (IIF(c2.ItemNumber = 8, c2.Item, ''), 'Email:', ''))
FROM #JSONish jn
    CROSS APPLY
(SELECT Cleaned = REPLACE (REPLACE (jn.Dataval, '''', ''), '"', '')) c1
    CROSS APPLY
(SELECT ss.* FROM dbo.DelimitedSplit8K (c1.Cleaned, ',') ss ) c2
GROUP BY jn.Id
ORDER BY jn.Id;

The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin

Phil Parkin SSC Guru Points: 246960 More actions · Answer 3

Obviously, this relies on Jeff Moden's splitter, but as you're on 2022, you can also use the native string_split() function.

The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin

Zond Sita SSC Veteran Points: 275 More actions · Answer 4

Thank you very much Phil. I am testing some more code for additions to my tables. will keep you post.

Thank you a Ton Phil.

Zond Sita SSC Veteran Points: 275 More actions · Answer 5

Good Morning Phil,

I have few issues I am encountering that is just my wrong data your solution is great.

Can I ask you a quick question, is there anyway can we just get all Values as column values and keys as headings of column names?

Please

I grateful to you, Phil

Thank you in advance

asita

Phil Parkin SSC Guru Points: 246960 More actions · Answer 6

If you are hoping for a dynamic solution which does not require you to type in the names of each key, that takes quite a chunk of coding and requires the use of dynamic SQL.

Alternatively, most reporting tools will do it for you. I was able to put the output of the inner query below (ie, the 'dta' CTE), and use PowerQuery to do all of the pivoting.

But if you are happy to type in all of the key names, you can use a variation of my first query

WITH dta
AS (SELECT jn.Id
          ,Ky = LEFT(c2.Item, CHARINDEX (':', c2.Item) - 1)
          ,Val = TRIM (RIGHT(c2.Item, LEN (c2.Item) - CHARINDEX (':', c2.Item)))
    FROM #JSONish jn
        CROSS APPLY
    (
        SELECT Cleaned = REPLACE (REPLACE (REPLACE (REPLACE (jn.Dataval, '''', ''), '"', ''), '{', ''), '}', '')
    ) c1
        CROSS APPLY
    (SELECT ss.* FROM dbo.DelimitedSplit8K (c1.Cleaned, ',') ss ) c2
    WHERE LEN (c2.Item) > 0)
SELECT d.Id
      ,cid = MAX (IIF(d.Ky = 'cid', d.Val, ''))
      ,id = MAX (IIF(d.Ky = 'id', d.Val, ''))
      ,Email = MAX (IIF(d.Ky = 'Email', d.Val, ''))
FROM dta d
GROUP BY d.Id
ORDER BY d.Id;

And here is the PQ I came up with:

let
    Source = Excel.CurrentWorkbook(){[Name="Table1"]}[Content],
    #"Changed Type" = Table.TransformColumnTypes(Source,{{"Id", Int64.Type}, {"Ky", type text}, {"Val", type text}}),
    #"Renamed Columns" = Table.RenameColumns(#"Changed Type",{{"Id", "Id1"}}),
    #"Pivoted Column" = Table.Pivot(#"Renamed Columns", List.Distinct(#"Renamed Columns"[Ky]), "Ky", "Val")
in
    #"Pivoted Column"

Which gives this

Pivot

The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin

Zond Sita SSC Veteran Points: 275 More actions · Answer 7

Thank you much Phil, I tweaked little hear and there your first code is working great. may I know what is this 8k 4k apart from String Split function? please

I owe you drink 🙂

God Bless you my friend.

Thanks Again

Best regards

Phil Parkin SSC Guru Points: 246960 More actions · Answer 8

DelimitedSplit8k is a highly optimised string splitter that was written in the days when there was nothing comparable built in to T-SQL. It has since become the de facto standard splitter for many people. For SQL Server 2022 and above, STRING_SPLIT() may be used instead.

You can find the code and supporting article here.

This reply was modified 9 months, 4 weeks ago by Phil Parkin.

The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin

Zond Sita SSC Veteran Points: 275 More actions · Answer 9

Got it Great! Thank you phil, one last question for this one

so here the value (example below record) "WCo:Underwriters at Lloyds, London" it is splitting the Underwriters at Lloyds , and london is going to another column value.

is there any fix that we try to avoid this in between spring splitting befroe next tag (:) just checking if there is any

{cd:430009532,PolId:c2e336-6448-4c9a-ac56-4f909cb38,PolNo:TFU-U,PolEff:2002-08-01,polExp:2020-08-01,stat:Active,LOB:Installation/Builder Risk (P),Prod:caiJei bigwood,PCo:comp Underwriters Inc,WCo:Underwriters at Lloyds, London, TOB:P,PGroup:Florida,ClaimPNum:(800) 334-1234,BMethod:Direct Bill},

SELECT
    jn.Id
,   cid = MAX (REPLACE (IIF(c2.ordinal = 3, c2.value, ''), 'cid:', ''))
,   Id = MAX (REPLACE (IIF(c2.ordinal = 4, c2.value, ''), 'Id:', ''))
,   Email = MAX (REPLACE (IIF(c2.ordinal = 8, c2.value, ''), 'Email:', ''))
FROM #JSONish jn
    CROSS APPLY
(SELECT Cleaned = REPLACE (REPLACE (jn.Dataval, '''', ''), '"', '')) c1
    CROSS APPLY
(SELECT ss.* FROM STRING_SPLIT(c1.Cleaned, ',', 1) ss) c2
GROUP BY jn.Id
ORDER BY jn.Id;

this is the code I am using just fyi. Thank you again.

Zond Sita SSC Veteran Points: 275 More actions · Answer 10

Dear phil,

I am thinking first replace comma with space or empty in the WCO string, then apply the logic

Thank you

Z Sita

Phil Parkin SSC Guru Points: 246960 More actions · Answer 11

Do you have any control over the format of the input file? If you could get them to use a delimiter which will not appear elsewhere in the text (| is a common one, I've also used ~ in the past), this problem would go away.

Otherwise, you have a problem, because (as per your initial post) some, but not all, of your data is 'text qualified'.

I can see the following solutions:

Make sure that all of your data is text qualified ("WCo:Underwriters at Lloyds, London" rather than WCo:Underwriters at Lloyds, London, for example). The code would have to be tweaked if this change were made, but it's doable.
Change the delimiter, as mentioned above
Change the data, so that it never contains a comma

The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin

Zond Sita SSC Veteran Points: 275 More actions · Answer 12

Thank you Phil.

There is no control in the history data which is what I am working on 🙁

So here what I am thinking is Replace commas in the WCO: value if any before we string split. then your code works and proved already

you think can we do any way?

 data_val2 = 
    STUFF(
        data_val,
        CHARINDEX('WCo:', data_val),
        CHARINDEX(',', data_val, CHARINDEX('WCo:', data_val)) - CHARINDEX('WCo:', data_val) + 1,
        REPLACE(
            SUBSTRING(
                data_value,
                CHARINDEX('WCo:', data_val),
                CHARINDEX(',', data_value, CHARINDEX('WCo:', data_val) + LEN('WCo:')) - CHARINDEX('WCo:', data_val)
            ),
            ',', ' '
        )
    )

from #my_temp
WHERE CHARINDEX('WCo:', data_val) > 0
  AND CHARINDEX(',', data_val, CHARINDEX('WCO:', data_val)) > 0;

but no luck yet.

Thanks Again

Zond