01 April 2021

💎SQL Reloaded: Processing JSON Files with Complex Structure in SQL Server 2016+

Unfortunately (or fortunately, for the challenge-searchers), not all JSON data files have a simple (matrix) structure, while the data might not even have a proper (readable) definition. It's the case of the unemployment data provided by the Cologne municipality (source). However with a language page translator and some small effort one can identify the proximate data definition:

Source FieldTarget FieldData Type
AM_ALO_INSG_AAALO_Totalint
AM_ALO_SGB2_AAALO_SGB2int
AM_ALO_UNTER25_AAALO_Under25int
AM_ALO_INSG_APALO_Total_Percfloat
AM_ALO_SGB2_APALO_SGB2_Percfloat
AM_ALO_UNTER25_APALO_Under25_Percfloat
AM_ALO_INSG_HAALO_Total_Histint
AM_ALO_SGB2_HAALO_SGB2_Histint
AM_ALO_UNTER25_HAALO_Under25_Histint
AM_ALO_INSG_HPALO_Total_HistPercfloat
AM_ALO_SGB2_HPALO_SGB2_HistPercfloat
AM_ALO_UNTER25_HPALO_Under25_HistPercfloat
AM_SVB_INSG_AASVB_Totalint
AM_SVB_MANN_AASVB_Menint
AM_SVB_FRAU_AASVB_Womenint
AM_SVB_DEUTSCH_AASVB_Germanint
AM_SVB_AUSLAND_AASVB_AUSLANDint
AM_SVB_U25J_AASVB_Under25Yoint
AM_SVB_UEBER55J_AASVB_Over55Yoint
AM_SVB_INSG_APSVB_Total_Percfloat
AM_SVB_MANN_APSVB_Men_Percfloat
AM_SVB_FRAU_APSVB_Women_Percfloat
AM_SVB_DEUTSCH_APSVB_German_Percfloat
AM_SVB_AUSLAND_APSVB_AUSLAND_Percfloat
AM_SVB_U25J_APSVB_Under25Yo_Percfloat
AM_SVB_UEBER55J_APSVB_Over55Yo_Percfloat
AM_SVB_INSG_HASVB_Total_Histint
AM_SVB_MANN_HASVB_Men_Histint
AM_SVB_FRAU_HASVB_Women_Histint
AM_SVB_DEUTSCH_HASVB_German_Histint
AM_SVB_AUSLAND_HASVB_AUSLAND_Histint
AM_SVB_U25J_HASVB_Under25Yo_Histint
AM_SVB_UEBER55J_HASVB_Over55Yo_Histint
AM_SVB_INSG_HPSVB_Total_HistPercfloat
AM_SVB_MANN_HPSVB_Men_HistPercfloat
AM_SVB_FRAU_HPSVB_Women_HistPercfloat
AM_SVB_DEUTSCH_HPSVB_German_HistPercfloat
AM_SVB_AUSLAND_HPSVB_AUSLAND_HistPercfloat
AM_SVB_U25J_HPSVB_Under25Yo_HistPercfloat
AM_SVB_UEBER55J_HPSVB_Over55Yo_HistPercfloat
SHAPE.AREASHAPE.AREAint
SHAPE.LENSHAPE.LENint

As previously stated (see post), it makes sense to build the logic over several iterations, making first sure that the references to file's columns were used correctly (observe the way the various elements were referenced in the queries):

SELECT DAT.ObjectId 
, DAT.Nummer
, DAT.Name
, DAT.ALO_Total
, DAT.ALO_SGB2
FROM OPENROWSET (BULK 'D:\data\Arbeitsmarkt Statistik Koeln Stadtteil.json',CODEPAGE='65001', SINGLE_CLOB)  as jsonfile 
     CROSS APPLY OPENJSON(BulkColumn,'$.features')
 WITH( 
	  ObjectId int '$.properties.OBJECTID'
	, Nummer int '$.properties.NUMMER'
	, Name nvarchar(max) '$.properties.NAME'
	, ALO_Total int '$.properties.AM_ALO_INSG_AA'
	, ALO_SGB2 int '$.properties.AM_ALO_SGB2_AA'
) AS DAT; 

Output (first 13 records):
ObjectIdNummerNameALO_TotalALO_SGB2
1211Godorf12192
2308Lövenich15683
3307Weiden552383
4306Junkersdorf305164
5309Widdersdorf20093
6404Vogelsang310208
7505Weidenpesch527351
8502Mauenheim190127
12207Hahnwald153
13213Meschenich644554

The logic seems to work, however the German umlauts aren't displayed as expected ('Lövenich', when it should have been 'Lövenich'). This is caused by the differences in character sets. An easy way to address this is to use a functions which does the conversion (see the dbo.ReplaceCodes2Umlauts UDF from an older post). 

By applying the function on the Names, adding the further columns and an INSERT clause, the query becomes:
 
-- importing the JSON file
SELECT DAT.ObjectId 
, DAT.Nummer
, dbo.ReplaceCodes2Umlauts(DAT.Name) Name
, DAT.ALO_Total
, DAT.ALO_SGB2
, DAT.ALO_Under25
, DAT.ALO_Total_Perc
, DAT.ALO_SGB2_Perc
, DAT.ALO_Under25_Perc
, DAT.ALO_Total_Hist
, DAT.ALO_SGB2_Hist
, DAT.ALO_Under25_Hist
, DAT.ALO_Total_HistPerc
, DAT.ALO_SGB2_HistPerc
, DAT.ALO_Under25_HistPerc
, DAT.SVB_Total
, DAT.SVB_Men
, DAT.SVB_Women
, DAT.SVB_German
, DAT.SVB_AUSLAND
, DAT.SVB_Under25Yo
, DAT.SVB_Over55Yo
, DAT.SVB_Total_Perc
, DAT.SVB_Men_Perc
, DAT.SVB_Women_Perc
, DAT.SVB_German_Perc
, DAT.SVB_AUSLAND_Perc
, DAT.SVB_Under25Yo_Perc
, DAT.SVB_Over55Yo_Perc
, DAT.SVB_Total_Hist
, DAT.SVB_Men_Hist
, DAT.SVB_Women_Hist
, DAT.SVB_German_Hist
, DAT.SVB_AUSLAND_Hist
, DAT.SVB_Under25Yo_Hist
, DAT.SVB_Over55Yo_Hist
, DAT.SVB_Total_HistPerc
, DAT.SVB_Men_HistPerc
, DAT.SVB_Women_HistPerc
, DAT.SVB_German_HistPerc
, DAT.SVB_AUSLAND_HistPerc
, DAT.SVB_Under25Yo_HistPerc
, DAT.SVB_Over55Yo_HistPerc
, DAT.Shape_Area 
, DAT.Shape_Len 
INTO dbo.Unemployment_Cologne
FROM OPENROWSET (BULK 'D:\data\Arbeitsmarkt Statistik Koeln Stadtteil.json',CODEPAGE='65001', SINGLE_CLOB)  as jsonfile 
     CROSS APPLY OPENJSON(BulkColumn,'$.features')
 WITH( 
	  ObjectId int '$.properties.OBJECTID'
	, Nummer int '$.properties.NUMMER'
	, Name nvarchar(max) '$.properties.NAME'
	, ALO_Total int '$.properties.AM_ALO_INSG_AA'
	, ALO_SGB2 int '$.properties.AM_ALO_SGB2_AA'
	, ALO_Under25 int '$.properties.AM_ALO_UNTER25_AA'
	, ALO_Total_Perc float '$.properties.AM_ALO_INSG_AP'
	, ALO_SGB2_Perc float '$.properties.AM_ALO_SGB2_AP'
	, ALO_Under25_Perc float '$.properties.AM_ALO_UNTER25_AP'
	, ALO_Total_Hist int '$.properties.AM_ALO_INSG_HA'
	, ALO_SGB2_Hist int '$.properties.AM_ALO_SGB2_HA'
	, ALO_Under25_Hist int '$.properties.AM_ALO_UNTER25_HA'
	, ALO_Total_HistPerc float '$.properties.AM_ALO_INSG_HP'
	, ALO_SGB2_HistPerc float '$.properties.AM_ALO_SGB2_HP'
	, ALO_Under25_HistPerc float '$.properties.AM_ALO_UNTER25_HP'
	, SVB_Total int '$.properties.AM_SVB_INSG_AA'
	, SVB_Men int '$.properties.AM_SVB_MANN_AA'
	, SVB_Women int '$.properties.AM_SVB_FRAU_AA'
	, SVB_German int '$.properties.AM_SVB_DEUTSCH_AA'
	, SVB_AUSLAND int '$.properties.AM_SVB_AUSLAND_AA'
	, SVB_Under25Yo int '$.properties.AM_SVB_U25J_AA'
	, SVB_Over55Yo int '$.properties.AM_SVB_UEBER55J_AA'
	, SVB_Total_Perc float '$.properties.AM_SVB_INSG_AP'
	, SVB_Men_Perc float '$.properties.AM_SVB_MANN_AP'
	, SVB_Women_Perc float '$.properties.AM_SVB_FRAU_AP'
	, SVB_German_Perc float '$.properties.AM_SVB_DEUTSCH_AP'
	, SVB_AUSLAND_Perc float '$.properties.AM_SVB_AUSLAND_AP'
	, SVB_Under25Yo_Perc float '$.properties.AM_SVB_U25J_AP'
	, SVB_Over55Yo_Perc float '$.properties.AM_SVB_UEBER55J_AP'
	, SVB_Total_Hist int '$.properties.AM_SVB_INSG_HA'
	, SVB_Men_Hist int '$.properties.AM_SVB_MANN_HA'
	, SVB_Women_Hist int '$.properties.AM_SVB_FRAU_HA'
	, SVB_German_Hist int '$.properties.AM_SVB_DEUTSCH_HA'
	, SVB_AUSLAND_Hist int '$.properties.AM_SVB_AUSLAND_HA'
	, SVB_Under25Yo_Hist int '$.properties.AM_SVB_U25J_HA'
	, SVB_Over55Yo_Hist int '$.properties.AM_SVB_UEBER55J_HA'
	, SVB_Total_HistPerc float '$.properties.AM_SVB_INSG_HP'
	, SVB_Men_HistPerc float '$.properties.AM_SVB_MANN_HP'
	, SVB_Women_HistPerc float '$.properties.AM_SVB_FRAU_HP'
	, SVB_German_HistPerc float '$.properties.AM_SVB_DEUTSCH_HP'
	, SVB_AUSLAND_HistPerc float '$.properties.AM_SVB_AUSLAND_HP'
	, SVB_Under25Yo_HistPerc float '$.properties.AM_SVB_U25J_HP'
	, SVB_Over55Yo_HistPerc float '$.properties.AM_SVB_UEBER55J_HP'
	, Shape_Area float '$.properties."SHAPE.AREA"'
	, Shape_Len float '$.properties."SHAPE.LEN"'
) AS DAT; 

Once the data made available, one can go on and discover the data and the relationships existing between the various columns. 

Happy coding!

No comments:

Related Posts Plugin for WordPress, Blogger...

About Me

My photo
Koeln, NRW, Germany
IT Professional with more than 24 years experience in IT in the area of full life-cycle of Web/Desktop/Database Applications Development, Software Engineering, Consultancy, Data Management, Data Quality, Data Migrations, Reporting, ERP implementations & support, Team/Project/IT Management, etc.