In this article, you will see few Examples to import REST API in SQL Server Table (Call JSON / XML SOAP API).  You will also learn how to load JSON Files into SQL Server Table using T-SQL Stored procedure code.  So let’s get ready for fun 🙂

JSON File format and REST API (i.e. RESTful API ) is becoming more and more popular each day. You may already have some data integration scenarios to read/write From REST API services in SQL Server or maybe API integration in other Apps such as Power BI, Tableau, SSIS, Informatica.

REST API integration can be challenging without the right set of tools. Even for highly skilled programmer sometimes its difficult to achieve the desired result.  But not to worry at the end of this article you will find out how easy it is to pull data from JSON / XML based web service using familiar SQL Query Language.  In the next few sections, you will learn how to call REST API inside SQL Server using familiar T-SQL code.

Steps to Import JSON / XML REST API in SQL Server

Here is the summary of steps to query REST API / JSON File in SQL Server (Using T-SQL code). For detailed steps read the full article and watch the video tutorial.

1. Install ZappySys ODBC PowerPack  from here (it contains drivers for JSON / XML API)
2. Create ODBC DSN for JSON / XML API Source
3. Configure ZappySys Data Gateway Service so SQL Server can use ZappySys Drivers in T-SQL code
4. Create and Configure Linked Server to connect to JSON / REST API Service
5. Write some T-SQL code in your stored procedure like below to import data from JSON / REST Service
   

   Example – Import JSON / XMLFiles / REST API Data into SQL Server Table ( Call API using T-SQL Stored Procedure)

Video Tutorial – Read from JSON / XML / REST API in SQL Server (T-SQL code)

If you want to save time and go quick then check out following video tutorial on calling JSON / REST API Service in SQL Server code. Steps The following video will be useful to configure your Linked Server and query REST API. It shows how to configure ODBC DSN, ZappySys Data Gateway Server (Used for Linked Server). To see other use cases of ZappySys Data Gateway click here

Requirements

1. A first requirement will be to SQL Server Database Engine Installed
2. The second requirement will be SSMS installed
3. Make sure to have ZappySys ODBC PowerPack installed.
4.  Another important requirement is the internet connection to connect to REST API information using a URL.

Step-By-Step : Load REST API in SQL Server

In next few sections, you will see step-by-step approach to consume REST API / JSON data in SQL Server using T-SQL code.

Configure ZappySys Data Gateway Service

To call external Drivers (e.g. ZappySys JSON/REST API Driver) in SQL Server we have to use Linked Server feature. We will use ZappySys Data Gateway Service along with SQL Linked Server to call ZappySys Drivers (e.g. JSON / XML API Drivers) using T-SQL Code in Stored Procedures.

Why use ZappySys Data Gateway Service?

Gateway Service is useful to call ZappySys driver from any other machine without installing ZappySys driver. For example your Unix server can consume API data by connecting to ZappySys Data Gateway (which can be running on Remote Windows OS). Many times you cannot access remote server but like to Query API data (Google SpreadSheet Use case). To learn more about other use cases click here.

Call REST API in SQL Server (T-SQL Code) using Linked Server

Optionally, use these lines of code to create the linked server using T-SQL instead of using the UI:

/* Add Linked Server called RESTAPI-SV (change if needed) to access REST API via Data Gateway - JSON/REST Driver */
/* customer_api is just an example data source name - created on Data gateway UI */

EXEC master.dbo.sp_addlinkedserver 
	@server = N'RESTAPI-SV', 
	@srvproduct=N'', 
--- For MSSQL 2012, 2014, 2016, 2017, and 2019 use below (SQL Server Native Client 11.0)---
    @provider=N'SQLNCLI11',
--- For MSSQL 2022 or higher use below (Microsoft OLE DB Driver for SQL Server)---
--- @provider=N'MSOLEDBSQL',
	@datasrc=N'localhost,5000', 
	@catalog=N'customer_api'

/* Add login for Linked Server - This is same user account and password which you define on Users tab on ZappySys Data Gateway Config Tool */
EXEC master.dbo.sp_addlinkedsrvlogin 
	@rmtsrvname=N'RESTAPI-SV',
	@useself=N'False',
	@locallogin=NULL,
	@rmtuser=N'admin',
	@rmtpassword='adminP@ss!23'
GO
--//OTHER IMPORTANT OPTIONS
EXEC master.dbo.sp_serveroption @server=N'RESTAPI-SV', @optname=N'rpc', @optvalue=N'true'
GO

EXEC master.dbo.sp_serveroption @server=N'RESTAPI-SV', @optname=N'rpc out', @optvalue=N'true'
GO

--Increase default timeout for long running query (default is 600 sec)
--EXEC master.dbo.sp_serveroption @server=N'RESTAPI-SV', @optname=N'query timeout', @optvalue=N'1200'
GO

--Below needed to support EXEC + INSERT (dynamic query)
EXEC master.dbo.sp_serveroption @server=N'RESTAPI-SV', @optname=N'remote proc transaction promotion', @optvalue=N'false'
GO

1. Now it is time to create some queries. Let’s run a simple query to REST API:
   

   select * from OPENQUERY([RESTAPI-SV] , 'SELECT * FROM value' )

2. The query will show all the data from REST API:
   

   REST API QUERY IN SQL SERVER

    

3. Also, the following query can be used when you need to use filters (Notice that we escaped single tick with double tick)

   select * from OPENQUERY([RESTAPI-SV] , 'SELECT * FROM value where ContactTitle=''Owner''' )

4. Store the data in a temporary table using the following query:

   select * INTO #tmpcustomer from OPENQUERY([RESTAPI-SV] , 'SELECT * FROM value' )

5. And now you can query and use the SQL table:

   SELECT * FROM #tmpcustomer t

6. You can also check the metadata, values and odata_next link:

   select * from OPENQUERY([RESTAPI-SV] , 'SELECT * FROM $' )

    

Call REST API in SQL Server (T-SQL Code) using Polybase (For SQL 2019 or higher)

How to use Dynamic SQL with parameters

So far we saw how to use OPENQUERY statement with a static SQL query. Now let’s look at the dynamic Query approach to Read REST API data and Load into SQL Table. For Dynamic SQL you have to use EXEC rather than OPENQUERY.

--MSDTC option must be off for INSERT + EXEC pattern (Find Linked Server Properties | Server Options | Enable Promotion of Distributed Transaction - [Set to false]
--OR-- run via code like below
--EXEC master.dbo.sp_serveroption 'MyLinkedServer', 'remote proc transaction promotion', false;

declare @tbl nvarchar(100) = 'value'
--Just output the query data (no insert)
EXEC('select col1,col2 from ' + @tbl) AT [MyLinkedServer]

--Load into table (dont output)
insert into _tmp_invoices(col1,col2) 
EXEC('select col1,col2 from ' + @tbl) AT [MyLinkedServer]

--We recommend using column names in the list rather than * (all) to avoid metdata issue
--Use Data Source > Preview UI > Table dropdown to generate column names

How to parameterize API calls in SQL Server

Now you know how to write simple SQL queries to fetch data from API. In real world you will have a need to pass parameters to your API. You can pass parameters to your API following ways.  Notice that how we have escaped single tick. In this example we have passed Header, Body and URL using Dynamic approach. In your case you may need to pass only one way (e.g. Pass parameter via URL only)

declare @url varchar(100)='http://httpbin.org/post' 
declare @body varchar(100)='{id:1,notes:"Line1\r\nLine2"}'
declare @header varchar(100)='Content-Type:text/plain||x-hdr1:AAA'

declare @sql varchar(1000)
declare @sqlForLs  varchar(1000) = 
'SELECT * FROM $
WITH  (
SRC='''+ replace(@url,'''', '''''')  +''' ,
METHOD=''POST'' ,
HEADER='''+ replace(@header,'''', '''''')  +''' ,
BODY='''+ replace(@body,'''', '''''')  +''' 
)'
 
set @sql = 'SELECT * into #tmp from OPENQUERY(LS,''' + replace(@sqlForLs,'''', '''''') + ''')'
print @sql

EXECUTE(@sql);

How to Import JSON File to SQL Server (Single or Multiple)

Not only you can query REST API in SQL Server but also you can query a JSON file. The steps are all the same except in one step. Instead of specifying a URL in the ZappySys ODBC driver (see Add the ZappySys driver in the Windows ODBC Administrator section, step 7) and specify the file path and set the format to Default:

Add properties JSON file

If you dont want to hard code file name in the DSN then you can also use following Query syntax to specify URL or File Path inside SQL Query

select * into #tmp from OPENQUERY([JSON-SV], 'SELECT * FROM value WITH (SRC=''C:\test\cust2012.json'')' )

How to load multiple JSON files into SQL Server Table

Let’s say that you have multiple JSON files in a folder (similar structure). The file names are customer_2012.json, customer_2013.json, customer_2014.json and they are all in the same folder. Is it possible to query all of them in SQL Server without any extra looping code?

The answer is yes. You just need to create the linked server to a single JSON file and then run this query (JSON-SV is the name of the linked server):

select * into #tmp from OPENQUERY([JSON-SV], 'SELECT * FROM value WITH (SRC=''C:\test\cust*.json'')' )

We are just using this code:

(SRC=''C:\test\cust*.json'')' )

SRC means source and it will search all the files with json extension starting with the word “cust”.

How to read and parse JSON string in T-SQL

In SQL Server 2016 and later, it is possible to read JSON data. However in older versions, it was not possible. The following example will show how to parse JSON data using a Linked Server named JSON-SV.

1. 1. We will first store JSON values in a SQL Server table in a column named MyData:
      

      create table #tmp(MyData nvarchar(max))
      
      insert into #tmp values(
      '{ 
      rows:[ 
      {id:1, name:"Sam''s Club"}, 
      {id:2, name:"Walmart"}, 
      {id:3, name: "Facebook"} 
      ]
      }')

   2. We just created a temporary table named tmp and in the column MyData, we stored JSON values. In the next step, we will store the values of the temporary table in a variable named json and store and parse the values en the variable sql:
      

      declare @json nvarchar(max)
      declare @sql nvarchar(max)
      
      select @json = MyData from #tmp
      
      
      set @sql='select * from OPENQUERY([JSON-SV] , ''SELECT * FROM rows WITH
      (DATA=@'''''+ Replace(@json ,'''','''''''''') +''''')'' );'
      print(@sql)
      execute(@sql)

   3. If everything is OK, you will be able to see the JSON values parsed in a nice table format:
      

      JSON values parsed

How to Call REST API from a SQL Stored Procedure

How can we call REST API using a stored procedure? We need to use dynamic queries to pass parameters. The following example will run a query where the city is specified in a variable.

1. We will create the stored procedure first:
   

   CREATE PROCEDURE getcity 
   @City NVARCHAR(50) 
   AS 
   DECLARE @sql NVARCHAR(500)='select * from OPENQUERY([RESTAPI-SV] , 
   ''SELECT * FROM value where City=''''' + @City + ''''''')' 
   EXECUTE(@sql)

2. The stored procedure name is getcity. We will call the RESTAPI-SV that calls the REST API and pass the parameter City in the query. We finally execute the query stored in the variable using the Execute statement.
3. Finally, we just invoke the stored procedure:
   

   EXECUTE getcity @City = N'Berlin'

   The query will display all the data with the city equal to Berlin.

How to POST data to REST API in SQL Server (Send data to URL)

The following example shows how to POST data to API service where data is coming from an embedded string in SQL Server using a Linked Server. I am assuming that you already created a linked server following previews steps explained before.

POST data to URL (embedded Body)

In SSMS run this query:

select * from OPENQUERY([RESTAPI-SV] , 'SELECT * FROM $ 
WITH (
 METHOD=''POST''
,HEADER=''Content-Type: application/json || x-hdr1: AAA''
,SRC=''http://httpbin.org/post''
,BODY=''{ id:2,  notes2:"Line1\r\nLine2" }''
)' 
)

The query will invoke the Linked Server for REST API, created before and use the POST Method with the embedded strings.

POST data to URL (read body from file)

Here is an example to POST data to URL where Body is read from Local file. Notice two additional attributes in below.  We set IsMultiPart=true and also Body has @ symbol before file path.

select * from OPENQUERY([RESTAPI-SV] , 'SELECT * FROM $ 
WITH (
 METHOD=''POST''
,HEADER=''Content-Type: application/json || x-hdr1: AAA''
,SRC=''http://httpbin.org/post''
,BODY=''@c:\data\api_call_body.json''
,IsMultiPart='True'
)' 
)

Avoiding errors with POST / PUT / DELETE  requests to create / modify /delete data

There will be a time you may get error when you issue POST requests because same request sent multiple times and duplicate records got created. To avoid such situation refer to this article (Read Handling POST request section)

Pagination Concepts for REST API / SOAP XML  in SQL Server

SELECT * FROM $
WITH(
SRC=@'https://zappysys.com/downloads/files/test/pagination_nextlink_inarray_1.json'
,NextUrlAttributeOrExpr = '$.nextlink'  --keep reading until this attribute is missing. If attribute name contains dot then use brackets like this $.['my.attr.name']
)

SELECT * FROM $
WITH (
SRC=@'https://zappysys.com/downloads/files/test/page-xml.aspx?page=1&mode=DetectBasedOnResponseStatusCode'
,PagingMode='ByUrlParameter'
,PagingByUrlAttributeName='page'
,PagingByUrlEndStrategy='DetectBasedOnResponseStatusCode'
,PagingByUrlCheckResponseStatusCode=401
,IncrementBy=1
)

SELECT * FROM $
WITH (
SRC=@'https://zappysys.com/downloads/files/test/cust-<%page%>.xml'
,PagingMode='ByUrlPath'
,PagingByUrlAttributeName='<%page%>'
,PagingByUrlEndStrategy='DetectBasedOnRecordCount'
,IncrementBy=1
)

SELECT * FROM $
LIMIT 25
WITH(
	 Src='https://wordpress.org/news/wp-json/wp/v2/categories?per_page=10'
	,PagingMode='ByResponseHeaderRfc5988'
	,WaitTimeMs='200' --//wait 200 ms after each request
)

Error Handling for REST API / SOAP XML in SQL Server

Sometimes errors occur... they just do and there is nothing you can do! Or can you? Actually, in ODBC PowerPack you can handle them in two ways.

METHOD 1 - Using Error Handling Options

When to use?

You may want to use them when your source is a resource located on the Internet; e.g. a file on a website, a file on an FTP server or just a plain API HTTP response. By default, when a remote server returns an error, data retrieval is stopped, an error is raised and no data is given back to you. This might not be always desirable.

Scenario 1

Imagine a scenario, that there is a web server which each day at 12 AM releases a new JSON file with that day's date as filename, e.g. http://www.some-server.com/data/2018-06-20.json. And, of course, you want to download it and use it daily in your Power BI report. But you have a problem: Power BI report data sources are refreshed each hour and you may get HTTP 404 status code (no file was found) when a file is not released yet. Which consequentially means other data sources won't be updated as well and you will see old and cached data on the report. That's where you could use Continue on any error or Continue when Url is invalid or missing (404 Errors) to avoid an error being raised and let other data sources to be updated.

Scenario 2

Another scenario is when you expect a web server to raise some kind of HTTP error when accessing a URL. You don't want ODBC Data Source to raise an error but instead, you want to get response data. That's where you can use Continue on any error or alike together with  Get response data on error to continue on an error and get the data:

METHOD 2 - Using Connection [Retry Settings]

Another scenario you may run into is a buggy web server. You ask it to give you some file or data and it, like a snotty kid, just doesn't give it to you! You have to ask twice or thrice before it does its job. If that's the case, you have to retry HTTP requests using Connection:

OAuth / HTTP Connection in SQL Server

If you need to authenticate or authorize your user to access a web resource, you will need to use one of the Connections:

• HTTP
• OAuth

HTTP Connection

Use HTTP Connection for simple Windows, Basic, NTLM or Kerberos authentication. Just fill in a username and a password and you are good to go!

You can also use HTTP Connection for more sophisticated authentication like:

• SOAP WSS (when accessing a SOAP WebService)
• Static Token / API Key (when need to pass an API key in HTTP header)
• Dynamic Token (same as Static Token method except that each time you need to log in and retrieve a fresh API key)
• JWT Token (As per RFC 7519)

OAuth

If you are trying to access REST API resource, it is a huge chance, you will need to use OAuth Connection. Read this article to understand how OAuth authentication and authorization works and how to use it (article originally was written for SSIS PowerPack, but the concepts and UI stay the same): 
https://zappysys.com/blog/rest-api-authentication-with-oauth-2-0-using-ssis/

Performance Tips for REST API / XML SOAP Calls

SELECT * FROM value
WITH(
	 Src='https://services.odata.org/V3/Northwind/Northwind.svc/Customers?$format=json&$filter=Country eq ''USA'''
	,DataFormat='Odata'
)

-- Slow query - Client-side filtering
SELECT * FROM value
WHERE Country ='USA'
WITH(
	 Src='https://services.odata.org/V3/Northwind/Northwind.svc/Customers?$format=json'
	,DataFormat='Odata'
)

SELECT * FROM $
WITH 
(  SRC='https://myhost.com/some-api'
  ,CachingMode='All'  --cache metadata and data rows both
  ,CacheStorage='File' --or Memory
  ,CacheFileLocation='c:\temp\myquery.cache'
  ,CacheEntryTtl=300 --cache for 300 seconds
)

SELECT * FROM $ 
LIMIT 10 --//add this just to test how fast you can get 10 rows
WITH(
  Filter='$.LargeArray[*]--FAST' --//Adding --FAST option turn on STREAM mode (large files)
 ,SRC='https://zappysys.com/downloads/files/test/large_file_100k_largearray_prop.json.gz'
 --,SRC='c:\data\large_file.json.gz'
 ,IncludeParentColumns='False'  --//This Must be OFF for STREAM mode (read very large files)
 ,FileCompressionType='GZip' --Zip or None (Zip format only available for Local files)
)

Other Considerations for Calling Web API in SQL Server

Calling XML SOAP Web Service in SQL Server

So far we have looked at examples to consume data using JSON driver. Now let’s look at an example, to call XML SOAP Web Service.

To call SOAP API you need to know Request XML Body Structure. If you are not sure how to create SOAP Request body then no worries. Check this article to learn how to generate SOAP Request body using the Free tool SoapUI. Basically, you have to use SoapUI to generate Request XML and after that, you can replace parameters as needed in the generated body.

What is SOAP Web Service?

If you are new to SOAP Web Service sometimes referred as XML Web Service then please read some concept about SOAP Web service standard from this link There are two important aspects in SOAP Web service.

1. Getting WSDL file or URL
2. Knowing exact Web Service URL

What is WSDL

In very simple term WSDL (often pronounced as whiz-dull) is nothing but a document which describes Service metadata (e.g. Functions you can call, Request parameters, response structure etc). Some service simply give you WSDL as xml file you can download on local machine and then analyze or sometimes you may get direct URL (e.g. http://api.mycompany.com/hr-soap-service/?wsdl )

Example SQL Query for SOAP API call using ZappySys XML Driver

Here is an example SQL query you can write to call SOAP API. If you not sure about many details then check next few sections on how to use XML Driver User Interface to build desired SQL query to POST data to XML SOAP Web Service without any coding.

SELECT * FROM $
WITH(
	 Src='http://www.holidaywebservice.com/HolidayService_v2/HolidayService2.asmx'
	,DataConnectionType='HTTP'
	,CredentialType='Basic' --OR SoapWss
	,SoapWssPasswordType='PasswordText'
	,UserName='myuser'
	,Password='pass$$w123'
	,Filter='$.soap:Envelope.soap:Body.GetHolidaysAvailableResponse.GetHolidaysAvailableResult.HolidayCode[*]'
	,ElementsToTreatAsArray='HolidayCode'	
	,RequestMethod='POST'	
	,Header='Content-Type: text/xml;charset=UTF-8 || SOAPAction: "http://www.holidaywebservice.com/HolidayService_v2/GetHolidaysAvailable"'
	,RequestData='
<soapenv:Envelope xmlns:soapenv="http://schemas.xmlsoap.org/soap/envelope/" xmlns:hol="http://www.holidaywebservice.com/HolidayService_v2/">
   <soapenv:Header/>
   <soapenv:Body>
      <hol:GetHolidaysAvailable>
         <!--type: Country - enumeration: [Canada,GreatBritain,IrelandNorthern,IrelandRepublicOf,Scotland,UnitedStates]-->
         <hol:countryCode>UnitedStates</hol:countryCode>
      </hol:GetHolidaysAvailable>
   </soapenv:Body>
</soapenv:Envelope>'
)

Now let's look at steps to create SQL query to call SOAP API. Later we will see how to generate code for your desired programming language (e.g. C# or SQL Server)

Video Tutorial - Introduction to SOAP Web Service and SoapUI tool

Before we dive into details about calling SOAP API using ZappySys XML Driver, lets first understand what is SOAP API and how to create SOAP requests using SoapUI tool. You will learn more about this process in the later section. The video contains some fragment about using SOAP API in SSIS but just ignore that part because we will be calling Soap API using ZappySys ODBC Driver rather than SSIS Components.  

Using SoapUI to test SOAP API call / Create Request Body XML

Assuming you have downloaded and installed SoapUI from here, now we are ready to use WSDL for your SOAP Web Service Calls. If you do not have WSDL file or URL handy then contact your API provider (sometimes you just have to add ?wsdl at the end of your Service URL to get WSDL so try that. Example: http://mycompany/myservice?wsdl ). If you don't know what is WSDL then in short, WSDL is Web service Description Language (i.e. XML file which describes your SOAP Service). WSDL helps to craft SOAP API request Body for ZappySys XML Driver. So Let's get started.

1. Open SoapUI and click SOAP button to create new SOAP Project
2. Enter WSDL URL or File Path of WSDLFor example WSDL for our sample service can be accessed via this URL

   http://www.dneonline.com/calculator.asmx?wsdl

   
   Create new SOAP API Project in SoapUI tool for SOAP API Testing
3. Once WSDL is loaded you will see possible operations you can call for your SOAP Web Service.
4. If your web service requires credentials then you have to configure it. There are two common credential types for public services (SOAP WSS or BASIC )
   1. To use SOAP WSS Credentials select request node and enter UserId, Password, and WSS-PasswordType (PasswordText or PasswordHash)
      
      Configure SOAP WSS Credentials for SoapUI (SOAP API Testing Tool)
   2. To use BASIC Auth Credentials select request node and double-click it. At the bottom click on Auth (Basic) and From Authorization dropdown click Add New and Select Basic.
      Configure Basic Authorization for SoapUI (SOAP API Testing Tool)
5. Now you can test your request first Double-click on the request node to open request editor.
6. Change necessary parameters, remove optional or unwanted parameters. If you want to regenerate request you can click on Recreate default request toolbar icon.
   Create SOAP Request XML (With Optional Parameters)
7. Once your SOAP Request XML is ready, Click the Play button in the toolbar to execute SOAP API Request and Response will appear in Right side panel. Test SOAP API using SoapUI Tool (Change Default XML Body / Parameters, Execute and See Response)

Create DSN using ZappySys XML Driver to call SOAP API

Once you have tested your SOAP API in SoapUI tool, we are ready to use ZappySys XML driver to call SOAP API in your preferred BI tool or Programming language.

1. First open ODBC Data Sources (search ODBC in your start menu or go under ZappySys > ODBC PowerPack > ODBC 64 bit)
2. Goto System DSN Tab (or User DSN which is not used by Service account)
3. Click Add and Select ZappySys XML Driver ZappySys ODBC Driver for XML / SOAP API
4. Configure API URL, Request Method and Request Body as below ZappySys XML Driver - Calling SOAP API - Configure URL, Method, Body
5. (This step is Optional) If your SOAP API requires credentials then Select Connection Type to HTTP and configure as below.
   ZappySys XML Driver - Configure SOAP WSS Credentials or Basic Authorization (Userid, Password)
6. Configure-Request Headers as below (You can get it from Request > Raw tab from SoapUI after you test the request by clicking the Play button) Configure SOAP API Request Headers - ZappySys XML Driver
7. Once credentials entered you can select Filter to extract data from the desired node. Make sure to select array node (see special icon) or select the node which contains all necessary columns if you don't have array node. Select Filter - Extract data from nested XML / SOAP API Response (Denormalize Hierarchy)
8. If prompted select yes to treat selected node as Array (This is helpful when you expect one or more record for selected node) Treat selected node as XML Array Option for SOAP API Response XML

Preview SOAP API Response / Generate SQL Code for SOAP API Call

Once you configure settings for XML Driver now you can preview data or generate example code for desired language (e.g. C#, Python, Java, SQL Server). Go to Preview tab and you will see default query generated based on settings you entered in previous sections. Attributes listed in WITH clause are optional. If you omit attribute in WITH clause it will use it from Properties tab.

Preview Data

Preview SOAP API Response in ZappySys XML Driver

Generate Code Option

Generate Example Code for ZappySys Driver

OPENQUERY vs EXEC AT (Handling Larger SQL Text)

So far we have seen examples of using OPENQUERY. It allows us to send pass-through query at remote server. The biggest limitation of OPENQUERY is it doesnt allow you to use variables inside SQL so often we have to use unpleasant looking dynamic SQL (Lots of tick, tick ….  and escape  hell). Well there is good news. With SQL 2005 and later you can use

EXEC( your_sql ) AT your_linked_server syntax

Disadvantage of EXEC AT is you cannot do SELECT INTO like OPENQUERY. Also you cannot perform JOIN like below in EXEC AT

SELECT a.* FROM OPENQUERY(ls_json,'select * from $') a 
JOIN  OPENQUERY(ls_json,'select * from $') b ON a.id=b.id

However you can always do INSERT INTO MyTable EXEC(…) AT LINKEDSRV. So table must exists when you do that way.

Here is how to use it. To use EXEC AT you must turn on RPC OUT option. Notice how we used variable in SQL to make it dynamic. This is much cleaner than previous approach we saw.

USE [master]
GO

EXEC master.dbo.sp_addlinkedserver @server = N'ls_Json', @srvproduct=N'', @provider=N'SQLNCLI', @datasrc=N'localhost,5000', @provstr=N'Network Library=DBMSSOCN;', @catalog=N'JsonApi'
EXEC master.dbo.sp_addlinkedsrvlogin @rmtsrvname=N'ls_Json',@useself=N'False',@locallogin=NULL,@rmtuser=N'tdsuser',@rmtpassword='########'
GO

EXEC sp_serveroption 'ls_Json', 'rpc out', true;  
go

declare @tbl varchar(100)='$'
EXEC('select * from ' + @tbl ) AT ls_Json

Here is the difference between OPENQUERY vs EXEC approaches.

SQL Server OPENQUERY vs EXEC for Linked Server – Handling Larger SQL Text (More than 8000 chars)

If you decide to use EXEC instead of OPENQUERY then make sure below setting is ON else you will get an error.

Under the Linked Server Options, Enable RPC and RPC Out and Disable Promotion of Distributed Transactions(MSDTC).

RPC OUT setting for EXEC AT statement in SQL Server Linked Server

Configure Firewall Settings

So far we have assumed that Gateway is running on the same machine as SQL Server. However there will be a case when ZappySys ODBC PowerPack is installed on a different machine than SQL Server.  In such case you may have to perform additional Firewall configurations. On most computers firewall settings wont allow outside traffic to ZappySys Data Gateway. In such case perform following steps to allow other machines to connect to Gateway.

1. Search for Windows Firewall Advanced Security in start menu.
2. Under Inbound Rules > Right click and click [New Rule] >> Click Next
3. Select Port on Rule Type >> Click Next
4. Click on TCP and enter port number under specified local port as 5000 (use different one if you changed Default port) >> Click Next
5. Select Profile (i.e. Private, Public) >> Click Next
6. Enter Rule name [i.e. ZappySys Data Gateway – Allow Inbound ] >> Click Next
7. Click OK to save the rule

ZappySys Datagateway – Allow Inbound Traffic on Port 5000 Firewall Rule

Known Issues

In this section we will discuss some known issues when using OPENQUERY / Data Gateway Connectivity

Error: The data is invalid

There will be a time when you start facing odd errors such as below. Possible

OLE DB provider "SQLNCLI11/MSOLEDBSQL" for linked server "Zs_Csv" returned message "Deferred prepare could not be completed.".
OLE DB provider "SQLNCLI11/MSOLEDBSQL" for linked server "Zs_Csv" returned message "Communication link failure".
Msg 13, Level 16, State 1, Line 0

Session Provider: The data is invalid.

Possible Cause:

There are few reasons for such error but below are two main reasons

• If query length is more than 2000 Characters long such as below then you might get such error
  

  SELECT * FROM OPENQUERY(LS, '--some really long text more than 2000 chars--')

• If query contains multiple OPENQUERY statements for JOIN / UNION like below then also it might fail (Gateway doesnt support parallel queries on a single connection – MARS compatibility issue).
  

  select a.id, b.name from OPENQUERY(LS, 'select * from tbl1') a join OPENQUERY(LS, 'select * from tbl2') b on a.id=b.id

Possible Fix:

There are few ways to fix above error based on reason why you getting this error (i.e. Query Length issue OR JOIN/UNION in the same statement)

• If your query has long SQL (more than 2000 chars ) then reduce SQL length using different techniques
  • e.g. use SELECT * FROM MyTable rather than SELECT col1,col2… FROM MyTable
  • Use Meta Option in WITH clause if you must use column name. (e.g. SELECT * FROM MyTable WITH(META=’c:\meta.txt’) this way you can define column in Meta file rather than SELECT query. Check this article
  • Consider using EXECT (….) AT [Linked_Server_name] option rather than OPENQUERY so you can use very long SQL (See next section on EXEC..AT usecase)
  • Consider using Virtual Table / Stored Proc to wrap long SQL so your call is very small (where usp_GetOrdersByYear is custom proc created on ZappySys Driver UI)
    

    SELECT * FROM OPENQUERY(LS, 'EXEC usp_GetOrdersByYear 2021')

• If your query uses JOIN  / UNION with multiple OPENQUERY in same SQL then use multiple Linked servers (one for each OPENQUERY clause) as below.
  

  select a.id, b.name from OPENQUERY(LS_1, 'select * from tbl1') a join OPENQUERY(LS_2, 'select * from tbl2') b on a.id=b.id

Error: Unable to begin a distributed transaction (When INSERT + EXEC used)

If you try to use EXEC statement and try to insert data into table (like below) you might get below error unless MSDTC option is turned off.

INSERT INTO MyTable EXEC('select * from tbl') AT MyLinkedServer

"Protocol error in TDS stream"
The operation could not be performed because OLE DB provider "SQLNCLI11" for linked server "ls_Json2" was unable to begin a distributed transaction.
--OR--
The operation could not be performed because OLE DB provider "MSOLEDBSQL" for linked server "ls_Json" was unable to begin a distributed transaction.

Solution:
Method-1: Go to linked server properties | Server Options | Enable Promotion of Distributed Transaction | Change to false (Default is true)
Now your try your INSERT with EXEC AT and it should work

Method-2: Run the below command if you dont want to use UI

EXEC master.dbo.sp_serveroption @server=N'My_Linked_Server', @optname=N'remote proc transaction promotion', @optvalue=N'false'

Error: Cannot use OPENQUERY with JOIN / UNION

When you perform JOIN  / UNION ALL on the same Linked Server it may fail to process sometimes because Datagateway doesnt support parallel query requests on the same connection. Workaround for that would be create multiple linked servers for the same datasource. See above section for the same workaround.

Error: Truncation errors due to data length mismatch

Many time you may receive truncation errors if Table column is length is less than actual column size from query column. To solve this issue use New version of Data gateway ( check – Use nvarchar(max) for string options – Found on General Tab)

Performance Tips

Now lets look at few performance tips

Use INSERT INTO rather than SELECT INTO to avoid extra META request

We discussed some Pros and Cons of OPENQUERY vs EXEC (…) AT in previous section. One obvious advantage of EXEC (….) AT is it reduces number of requests to driver (It sends pass through query). With EXEC you cannot load data dynamically like SELECT INTO tmp FROM OPENQUERY. Table must exist before hand if you use EXEC.

INSERT INTO tmp_API_Report_Load(col1,col2)
EXEC('select col1,col2 from some_api_table') AT [API-LINKED-SERVER]
--OR--
INSERT INTO tmp_API_Report_Load(col1,col2)
select col1,col2 from OPENQUERY([API-LINKED-SERVER], 'select col1,col2 from some_api_table')

Advantage of this method is your Query speed will increase because system calls API only once when you call EXEC AT. In OPENROWSET it needs to call above query twice (Once to obtain metadata and once to get data).

Use Cached Metadata if possible

By default, most of the SQL sent to Data gateway need to invoke two phases. First, get metadata and second fetch data. You can bypass meta API call by supplying static metadata. Use META property in WITH clause as per this article to speed up your SQL.

Conclusion

To conclude, we can say that using ZappySys JSON / XML Drivers we can query JSON / XML Files and call REST API or SOAP Web Service using familiar SQL Language. All you need to do is install ZappySys ODBC Drivers, configure it in the ODBC DSN and use ZappySys Gateway as a Proxy. Finally, you add a Linked Server in SQL Server to connect to ZappySys Data Gateway. Feel free to download  ZappySys Drivers here. You can contact ZappySys Team here if you need any API integration help.

References

• ZappySys ODBC PowerPack online help
• JSON data in SQL Server
• Linked Servers (Database Engine)

Keywords: Call REST API, SQL SERVER, JSON

The post How to import REST API in SQL Server (Call JSON / XML SOAP Service) appeared first on ZappySys Blog.