C#'s foreach ruined my afternoon

Published Tue 21 March 2017 in Programming

"Forest Fire" by CIFOR is licensed under CC BY-NC-ND 2.0

The other afternoon I ran into some nightmarish debugging with the following code:

private static void StartThreads()
{
    var values = new List<int>() { 1, 2, 3 };
    var threads = new List<Thread>();

    foreach (var value in values)
    {
        Thread t = new Thread(() => Run(value));
        threads.Add(t);
        t.Start();
    }

    // Wait for all threads to complete
    foreach (Thread t in threads)
        t.Join();
}

private static void Run(int value)
{
    Console.Write(value.ToString());
}

(I know, I know, I wish I could be using TPL but in this case I couldn't)

On my local machine, the code above ran and gave me my expected console output of 123 (your results may vary depending on what order the threads run in).

When I ran this code on my server however, the output was 333.

begin pulling out hair

Long story short, after a couple hours of investigation I figured out that the way a foreach loop works under the hood in C# ≥ 5.0, which is what I run on my local machine, works differently than a foreach loop in C# < 5.0, which is what I had on my server.

From the C# 4.0 spec, a foreach loop is really a while loop behind the scenes, meaning the code above really translates into something like this:

private static void StartThreads()
{
    var values = new List<int>() { 1, 2, 3 };
    var threads = new List<Thread>();
    IEnumerator<int> e = ((IEnumerable<int>)values).GetEnumerator();

    try
    {
        int v; // DECLARED OUTSIDE OF THE LOOP!!!
        while (e.MoveNext())
        {
            v = (int)(int)e.Current;
            Thread t = new Thread(() => Run(v));
            threads.Add(t);
            t.Start();
        }
    }
    finally
    {
        if (e != null) ((IDisposable)e).Dispose();
    }

    // Wait for all threads to complete
    foreach (Thread t in threads)
        t.Join();
}

The important thing to note in the above code is that int v gets declared outside of the while loop.

In the C# 5.0 spec, that int v gets declared inside the loop (causing it to get recreated with every iteration):

private static void StartThreads()
{
    var values = new List<int>() { 1, 2, 3 };
    var threads = new List<Thread>();
    IEnumerator<int> e = ((IEnumerable<int>)values).GetEnumerator();

    try
    {
        while (e.MoveNext())
        {
            int v; // C# 5.0 DECLARED INSIDE THE LOOP
            v = (int)(int)e.Current;
            Thread t = new Thread(() => Run(v));
            threads.Add(t);
            t.Start();
        }
    }
    finally
    {
        if (e != null) ((IDisposable)e).Dispose();
    }

    // Wait for all threads to complete
    foreach (Thread t in threads)
        t.Join();
}

Because my local machine and server were running different versions of .NET, the same exact code was producing totally different results.

Eventually I found Eric Lippert's article about the matter. Since I'm still fairly new to the world of .NET, I wasn't around for the big debate that took place in his comment's section regarding which should be the correct implementation. However, it is interesting to note that the C# devs decided to switch the logic on how the foreach loop operates so late in the game.

My eventual workaround for the .NET 3.5/C# 4.0 server was to assign the int to a newly created variable inside the foreach:

foreach (var value in values)
{
    var tempValue = value; // THE FIX
    Thread t = new Thread(() => Run(tempValue));
    threads.Add(t);
    t.Start();
}

As frustrating it may be to debug problems like this, it is nice to learn a little bit more of the language's history and idiosyncrasies.

When Is It Appropriate To Store JSON in SQL Server?

Published Tue 14 March 2017 in SQL > Development > JSON

Who needs a relational database when everything can be stored in a JSON string?

Every once in a while I hear of some technologist say that relational databases are dead; instead, a non-table based NoSQL storage format is the way of the future. SQL Server 2016 introduced JSON functionality, making it possible for some "non-SQL" data storage to make its way into the traditionally tabled-based SQL Server.

Does this mean all data in SQL Server going forward should be stored in long JSON strings? No, that would be a terrible idea. There are instances when storing JSON in SQL Server is a good choice though. In this post I want to create recommendations for when data should be stored as JSON and when it shouldn't.

Databases Should Not Be Entirely Comprised Of JSON

The screenshot below is an example of what I think some developers would do if they were given free reign in SQL Server 2016:

bb208-1zzw7xinj5wmotojplta2lw

Here we have an application database ("InventoryApp") that consists of only a single table ("dbo.Data") with three JSON NVARCHAR(MAX) columns to represent all of the data required by the app. Relationships exist between Sales, Purchases, and Customers but these are not defined on the database side.

If you are from the world of relational-SQL, you might not believe that anyone would design such a database structure. Believe me though, this is a realistic scenario. Entire companies (eg. Firebase: https://firebase.google.com/) build their services around abstracting the database layer away from developers, essentially storing entire tables or databases in large JSON strings.

Many developers like storing data this way because it is easy to deserialize JSON strings into objects in their programming languages to use in their apps. They like the fact that with JSON they can have an infinitely changing storage schema (just add new keys, values, and arrays!) so if they need a new field for their app, they can just add it in, serialize the object to a JSON string, and store it again in the database.

Obviously, going completely "NoSQL" might make short term development easier/quicker, but using SQL Server 2016 to only store data this way is a travesty: there's no way to use many of SQL Server's amazing performance, schema definition and validation, and security features.

So when is it appropriate to store JSON in SQL Server?

Appropriate Use Case #1: Error Logging

Errors happen. When they do, it's nice to be able to go back and look at the error message to see what happened.

The problem is that the structure of error messages isn't always consistent. Sometimes only the value of a single property will help identify the cause of failure. Other times, something more complex fails and it would be nice to have all of the values of a complex object available to make troubleshooting easier.

This is where JSON steps in: in most programming languages, it is easy to convert error messages and run time values to a JSON object on error. And since error messages and data values change in structure depending on where they occur, it's easy to dynamically turn any type of object into JSON data.

This data is perfect to store in SQL to be looked at later. None of these ideas are new — nvarchar(max) has been in SQL for a while now, and so programmers everywhere have been storing error information in that datatype.

With SQL Server 2016, it is now easier to examine and parse the error information directly in SQL Server Management Studio with the variety of JSON parsing functions available. No longer do programmers have to copy the code into some different tool — they can easily do it in SSMS.

Appropriate Use Case #2: Piloting Ideas

Most large workplaces have controls in place that prevent developers from making changes in production. In general this is a Good Idea™.

However, controls are sometimes too restrictive. For example, due to security restrictions, lack of server space, company politics, etc… developers are sometimes stuck developing in production. It's an unfortunate fact of life. In those scenarios, developers have to go through hell if they have to elevate each database structure change every time they want to test something in production.

JSON to the rescue! An nvarchar(max) column in a table can have its JSON data be easily added to and modified to fit more data than it was originally intended to hold. All without any database structure change requests.

Now this is not an ideal situation. In fact, it's a scenario that can add a lot of technical debt to the application long-term if not planned for.

However, if a "flexible" JSON column is built with eventual conversion to a traditional table structure in mind from the start, it's actually simple for a developer to transition an entirely JSON storage structure to a relational format later on. They key here is that the developer needs to have this conversion planned from day one.

Appropriate Use Case #3: Non-Analytical Data

Analytical data is SQL Server's bread and butter. Need to store lots of data and be able to query against it all day long? No problem, there are a plethora of performance tuning options to make your queries run fast and efficiently.

However, sometimes not all data needs to be analyzed. Often an app might need to save some session data to a database temporarily — why bother creating all of the maintenance overhead of strict database schemas if the data will never be queried for analytical purposes? Another example might be a website's dynamically created user profile settings. You can build normalized table(s) to store all of that data, but then you will be writing programming logic to normalize and denormalize your data out of the app.

If this data will not have to be searched, then why bother adding all of the overhead? Keep it in JSON and be done with.

The Ultimate SQL Server JSON Cheat Sheet

Published Tue 07 March 2017 in SQL > Development > JSON

f23b7-1bpskynllsyfmagyrr-svwa

This post is a reference of my examples for processing JSON data in SQL Server. For more detailed explanations of these functions, please see my post series on JSON in SQL Server 2016:

Additionally, the complete reference for SQL JSON handling can be found at MSDN: https://msdn.microsoft.com/en-us/library/dn921897.aspx

Parsing JSON

Getting string JSON data into a SQL readable form.

ISJSON()

Checks to see if the input string is valid JSON.

SELECT ISJSON('{ "Color" : "Blue" }') -- Returns 1, valid
-- Output: 1

SELECT ISJSON('{ "Color" : Blue }') -- Returns 0, invalid, missing quotes
-- Output: 0

SELECT ISJSON('{ "Number" : 1 }') -- Returns 1, valid, numbers are allowed
-- Output: 1

SELECT ISJSON('{ "PurchaseDate" : "2015-08-18T00:00:00.000Z" }') -- Returns 1, valid, dates are just strings in ISO 8601 date format https://en.wikipedia.org/wiki/ISO_8601
-- Output: 1

SELECT ISJSON('{ "PurchaseDate" : 2015-08-18 }') -- Returns 0, invalid
-- Output: 0

JSON_VALUE()

Extracts a specific scalar string value from a JSON string using JSON path expressions.

-- See https://gist.github.com/bertwagner/356bf47732b9e35d2156daa943e049e9 for a formatted version of this JSON
DECLARE @garage nvarchar(1000) = N'{ "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GL" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }] }'

SELECT JSON_VALUE(@garage, '$.Cars[0].Make') -- Return the make of the first car in our array
-- Output: Volkswagen

SELECT CAST(JSON_VALUE(@garage, '$.Cars[0].PurchaseDate') as datetime2) -- Return the Purchase Date of the first car in our array and convert it into a DateTime2 datatype
-- Output: 2006-10-05 00:00:00.0000000

SELECT JSON_VALUE(@garage, '$.Cars') -- This returns NULL because the values of Cars is an array instead of a simple object
-- Output: NULL

SELECT JSON_VALUE(@garage, '$.Cars[1].Model') -- This is also invalid because JSON_VALUE cannot return an array...only scalar values allowed!
-- Output: NULL

SELECT JSON_VALUE(@garage, '$.Cars[1].Model.Base') -- Much better
-- Output: Impreza

Strict vs. Lax mode

If the JSON path cannot be found, determines if the function should return a NULL or an error message.

-- Lax (default: function will return an error if invalid JSON path specified
SELECT JSON_VALUE('{ "Color" : "Red" }', '$.Shape') --lax is the default, so you don't need to be explicitly state it
-- Output: NULL

SELECT JSON_VALUE('{ "Color" : "Red" }', 'lax $.Shape')
-- Output: NULL

-- Strict: function will return an error if invalid JSON path specified
SELECT JSON_VALUE('{ "Color" : "Red" }', 'strict $.Shape')
-- Output: Property cannot be found on the specified JSON path.

JSON_QUERY()

Returns a JSON fragment for the specified JSON path.

-- See https://gist.github.com/bertwagner/356bf47732b9e35d2156daa943e049e9 for a formatted version of this JSON
DECLARE @garage nvarchar(1000) = N'{ "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GL" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }] }'

-- This returns NULL because the values of Cars is an array instead of a simple object
SELECT JSON_VALUE(@garage, '$.Cars') 
-- Output: NULL

-- Using JSON_QUERY() however returns the JSON string representation of our array object
SELECT JSON_QUERY(@garage, '$.Cars') 
-- Output: [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GL" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }]

-- This instance of JSON_VALUE() correctly returns a singular scalar value
SELECT JSON_VALUE(@garage, '$.Cars[0].Make')
-- Output: Volkswagen

-- Using JSON_QUERY will not work for returning scalar values - it only will return JSON strings for complex objects
SELECT JSON_QUERY(@garage, '$.Cars[0].Make')
-- Output: NULL

This is useful to help filter an array and then extract values with JSON_VALUE():

-- See https://gist.github.com/bertwagner/356bf47732b9e35d2156daa943e049e9 for a formatted version of this JSON
DECLARE @garage nvarchar(1000) = N'{ "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GL" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }] }'

-- We use JSON_QUERY to get the JSON representation of the Cars array
SELECT JSON_QUERY(@garage, '$.Cars')
-- Output: [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GL" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }]

-- If we combine it with JSON_VALUE we can then pull out specific scalar values
SELECT JSON_VALUE(JSON_QUERY(@garage, '$.Cars') , '$[0].Make')
-- Output: Volkswagen

OPEN_JSON()

Returns a SQL result set for the specified JSON path. The result set includes columns identifying the datatypes of the parsed data.

-- See https://gist.github.com/bertwagner/356bf47732b9e35d2156daa943e049e9 for a formatted version of this JSON
DECLARE @garage nvarchar(1000) = N'{ "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GL" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }] }'

SELECT * FROM OPENJSON(@garage, '$.Cars') -- Displaying the values of our "Cars" array.  We additionally get the order of the JSON objects outputted in the "key" column and the JSON object datatype in the "type" column
/* Output:
key    value                                                                                                                                type
------ ------------------------------------------------------------------------------------------------------------------------------------ ----
0      { "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GL" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }        5
1      { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }    5
*/

SELECT * FROM OPENJSON(@garage, '$.Cars[0]') -- Specifying the first element in our JSON array.  JSON arrays are zero-index based
/* Output:
key              value                                 type
---------------- ------------------------------------- ----
Make             Volkswagen                            1
Model            { "Base": "Golf", "Trim": "GL" }      5
Year             2003                                  2
PurchaseDate     2006-10-05T00:00:00.000Z              1
*/

SELECT * FROM OPENJSON(@garage, '$.Cars[0].Model') -- Pulling the Model property from the first element in our Cars array
/* Output:
key     value   type
------- ------- ----
Base    Golf    1
Trim    GL      1
*/

-- See https://gist.github.com/bertwagner/356bf47732b9e35d2156daa943e049e9 for a formatted version of this JSON
DECLARE @garage nvarchar(1000) = N'{ "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GL" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }] }'

-- Here we retrieve the Make of each vehicle in our Cars array
SELECT JSON_VALUE(value, '$.Make') FROM OPENJSON(@garage, '$.Cars') 
/* Output: 
------------
Volkswagen
Subaru
*/ 

-- Parsing and converting some JSON dates to SQL DateTime2
SELECT CAST(JSON_VALUE(value, '$.PurchaseDate') as datetime2) FROM OPENJSON(@garage, '$.Cars') 
/* Output: 
---------------------------
2006-10-05 00:00:00.0000000
2015-08-18 00:00:00.0000000
*/ 

-- We can also format the output schema of a JSON string using the WITH option.  This is especially cool because we can bring up values from sub-arrays (see Model.Base and Model.Trim) to our top-level row result
SELECT * FROM OPENJSON(@garage, '$.Cars')
 WITH (Make varchar(20) 'strict $.Make',  
       ModelBase nvarchar(100) '$.Model.Base',
       ModelTrim nvarchar(100) '$.Model.Trim',
        Year int '$.Year',  
       PurchaseDate datetime2 '$.PurchaseDate') 
/* Output: 
Make           ModelBase   Year        PurchaseDate
-------------- ----------- ----------- ---------------------------
Volkswagen     Golf        2003        2006-10-05 00:00:00.0000000
Subaru         Impreza     2016        2015-08-18 00:00:00.0000000
*/

Creating JSON

Creating JSON data from either strings or result sets.

FOR JSON AUTO

Automatically creates a JSON string from a SELECT statement. Quick and dirty.

-- Create our table with test data
DROP TABLE IF EXISTS ##Garage;
CREATE TABLE ##Garage
(
    Id int IDENTITY(1,1),
    Make varchar(100),
    BaseModel varchar(50),
    Trim varchar(50),
    Year int,
    PurchaseDate datetime2
);
INSERT INTO ##Garage VALUES ('Volkswagen', 'Golf', 'GL', 2003, '2006-10-05');
INSERT INTO ##Garage VALUES ('Subaru', 'Impreza', 'Premium', 2016, '2015-08-18');

-- Take a look at our data
SELECT * FROM ##Garage;

-- AUTO will format a result into JSON following the same structure of the result set
SELECT Make, BaseModel, Trim, Year, PurchaseDate
FROM ##Garage
FOR JSON AUTO;
-- Output: [{"Make":"Volkswagen","BaseModel":"Golf","Trim":"GL","Year":2003,"PurchaseDate":"2006-10-05T00:00:00"},{"Make":"Subaru","BaseModel":"Impreza","Trim":"Premium","Year":2016,"PurchaseDate":"2015-08-18T00:00:00"}]

-- Using aliases will rename JSON keys
SELECT Make AS [CarMake] 
FROM ##Garage 
FOR JSON AUTO;
-- Output: [{"CarMake":"Volkswagen"},{"CarMake":"Subaru"}]

-- Any joined tables will get created as nested JSON objects.  The alias of the joined tables becomes the name of the JSON key
SELECT g1.Make,  Model.BaseModel as Base, Model.Trim, g1.Year, g1.PurchaseDate
FROM ##Garage g1
INNER JOIN ##Garage Model on g1.Id = Model.Id
FOR JSON AUTO;
-- Output: [{"Make":"Volkswagen","Year":2003,"PurchaseDate":"2006-10-05T00:00:00","Model":[{"Base":"Golf","Trim":"GL"}]},{"Make":"Subaru","Year":2016,"PurchaseDate":"2015-08-18T00:00:00","Model":[{"Base":"Impreza","Trim":"Premium"}]}]

-- Finally we can encapsulate our entire JSON result in a parent element by specifiying the ROOT option
SELECT Make, BaseModel, Trim, Year, PurchaseDate
FROM ##Garage
FOR JSON AUTO, ROOT('Cars');
-- Output: {"Cars":[{"Make":"Volkswagen","BaseModel":"Golf","Trim":"GL","Year":2003,"PurchaseDate":"2006-10-05T00:00:00"},{"Make":"Subaru","BaseModel":"Impreza","Trim":"Premium","Year":2016,"PurchaseDate":"2015-08-18T00:00:00"}]}

FOR JSON PATH

Formats a SQL query into a JSON string, allowing the user to define structure and formatting.

-- PATH will format a result using dot syntax in the column aliases.  Here's an example with just default column names
SELECT Make, BaseModel, Trim, Year, PurchaseDate
FROM ##Garage
FOR JSON PATH, ROOT('Cars');
-- Output: {"Cars":[{"Make":"Volkswagen","BaseModel":"Golf","Trim":"GL","Year":2003,"PurchaseDate":"2006-10-05T00:00:00"},{"Make":"Subaru","BaseModel":"Impreza","Trim":"Premium","Year":2016,"PurchaseDate":"2015-08-18T00:00:00"}]}

-- And here is the same example, just assigning aliases to define JSON nested structure
SELECT Make, BaseModel as [Model.Base], Trim AS [Model.Trim], Year, PurchaseDate
FROM ##Garage
FOR JSON PATH, ROOT('Cars');
-- Output: {"Cars":[{"Make":"Volkswagen","Model":{"Base":"Golf","Trim":"GL"},"Year":2003,"PurchaseDate":"2006-10-05T00:00:00"},{"Make":"Subaru","Model":{"Base":"Impreza","Trim":"Premium"},"Year":2016,"PurchaseDate":"2015-08-18T00:00:00"}]}

-- We can actually go multiple levels deep with this type of alias dot notation nesting
SELECT Make, BaseModel as [Model.Base], Trim AS [Model.Trim], 'White' AS [Model.Color.Exterior], 'Black' AS [Model.Color.Interior], Year, PurchaseDate
FROM ##Garage
FOR JSON PATH, ROOT('Cars');
-- Output: {"Cars":[{"Make":"Volkswagen","Model":{"Base":"Golf","Trim":"GL","Color":{"Exterior":"White","Interior":"Black"}},"Year":2003,"PurchaseDate":"2006-10-05T00:00:00"},{"Make":"Subaru","Model":{"Base":"Impreza","Trim":"Premium","Color":{"Exterior":"White","Interior":"Black"}},"Year":2016,"PurchaseDate":"2015-08-18T00:00:00"}]}

-- Concatenating data rows with UNION or UNION ALL just adds the row as a new element as part of the JSON array
SELECT Make,  BaseModel AS [Model.Base], Trim AS [Model.Trim], Year, PurchaseDate
FROM ##Garage WHERE Id = 1
UNION ALL
SELECT Make,  BaseModel, Trim, Year, PurchaseDate
FROM ##Garage WHERE Id = 2
FOR JSON PATH, ROOT('Cars');
-- Output: {"Cars":[{"Make":"Volkswagen","Model":{"Base":"Golf","Trim":"GL"},"Year":2003,"PurchaseDate":"2006-10-05T00:00:00"},{"Make":"Subaru","Model":{"Base":"Impreza","Trim":"Premium"},"Year":2016,"PurchaseDate":"2015-08-18T00:00:00"}]}

-- We can even include our FOR JSON in our SELECT statement to generate JSON strings for each row of our result set
SELECT g1.*, (SELECT Make, BaseModel AS [Model.Base], Trim AS [Model.Trim], Year, PurchaseDate FROM ##Garage g2 WHERE g2.Id = g1.Id FOR JSON PATH, ROOT('Cars')) AS [Json]
FROM ##Garage g1
/* Output: 
Id  Make          BaseModel    Trim      Year    PurchaseDate                Json
--- ------------- ------------ --------- ------- --------------------------- --------------------------------------------------------------------------------------------------------------------------
1   Volkswagen    Golf         GL        2003    2006-10-05 00:00:00.0000000 {"Cars":[{"Make":"Volkswagen","Model":{"Base":"Golf","Trim":"GL"},"Year":2003,"PurchaseDate":"2006-10-05T00:00:00"}]}
2   Subaru        Impreza      Premium   2016    2015-08-18 00:00:00.0000000 {"Cars":[{"Make":"Subaru","Model":{"Base":"Impreza","Trim":"Premium"},"Year":2016,"PurchaseDate":"2015-08-18T00:00:00"}]}
*/

Modifying JSON

Updating, adding to, and deleting from JSON data.

JSON_MODIFY()

Allows the user to update properties and values, add properties and values, and delete properties and values (the delete is unintuitive, see below).

Modify:

-- See https://gist.github.com/bertwagner/356bf47732b9e35d2156daa943e049e9 for a formatted version of this JSON
DECLARE @garage nvarchar(1000) = N'{ "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GL" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }] }'

-- I upgraded some features in my Volkswagen recently, technically making it equivalent to a "GLI" instead of a "GL".  
-- Let's update our JSON using JSON_MODIFY:
SET @garage = JSON_MODIFY(@garage, '$.Cars[0].Model.Trim', 'GLI')
SELECT @garage
-- Output: { "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GLI" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }] }

Add:

-- See https://gist.github.com/bertwagner/356bf47732b9e35d2156daa943e049e9 for a formatted version of this JSON
DECLARE @garage nvarchar(1000) = N'{ "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GLI" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }] }'

-- I decided to sell my Golf.  Let's add a new "SellDate" property to the JSON saying when I sold my Volkswagen.
-- If we use strict mode, you'll see we can't add SellDate because the key never existed before
--SELECT JSON_MODIFY(@garage, 'append strict $.Cars[0].SellDate', '2017-02-17T00:00:00.000Z')
-- Output: Property cannot be found on the specified JSON path.

-- However, in lax mode (default), we have no problem adding the SellDate
SELECT JSON_MODIFY(@garage, 'append lax $.Cars[0].SellDate', '2017-02-17T00:00:00.000Z')
-- Output: { "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GLI" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" ,"SellDate":["2017-02-17T00:00:00.000Z"]}, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" }] }

-- After selling my Golf, I bought another car a few days later: A new Volkswagen Golf GTI.  Let's add it to our garge:
-- Note the use of JSON_QUERY; this is so our string is interpreted as a JSON object instead of a plain old string
SET @garage = JSON_MODIFY(@garage, 'append $.Cars', JSON_QUERY('{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GTI" }, "Year": 2017, "PurchaseDate": "2017-02-19T00:00:00.000Z" }'))
SELECT @garage;
-- Output: { "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GLI" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" },{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GTI" }, "Year": 2017, "PurchaseDate": "2017-02-19T00:00:00.000Z" }] }

Delete property:

DECLARE @garage nvarchar(1000) = N'{ "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GLI" }, "Year": 2003, "PurchaseDate": "2006-10-05T00:00:00.000Z", "SellDate" : "2017-02-17T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" },{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GTI" }, "Year": 2017, "PurchaseDate": "2017-02-19T00:00:00.000Z" }] }'

-- Let's remove the PurchaseDate property on my original Volkswagen Golf since it's not relevant anymore:
SET @garage = JSON_MODIFY(@garage, '$.Cars[0].PurchaseDate', NULL)
SELECT @garage
-- Output: { "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GLI" }, "Year": 2003, "SellDate" : "2017-02-17T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" },{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GTI" }, "Year": 2017, "PurchaseDate": "2017-02-19T00:00:00.000Z" }] }

Delete from array (this is not intuitive, see my Microsoft Connect item to fix this: https://connect.microsoft.com/SQLServer/feedback/details/3120404/sql-modify-json-null-delete-is-not-consistent-between-properties-and-arrays)

DECLARE @garage nvarchar(1000) = N'{ "Cars": [{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GLI" }, "Year": 2003, "SellDate" : "2017-02-17T00:00:00.000Z" }, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" },{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GTI" }, "Year": 2017, "PurchaseDate": "2017-02-19T00:00:00.000Z" }] }'

-- I realize it's not worth keeping the original Volkswagen in my @garage data any longer, so let's completely remove it.
-- Note, if we use NULL as per the MSDN documentation, we don't actually remove the first car element of the array - it just gets replaced with NULL
-- This is problematic if we expect the indexes of our array to shift by -1.
SELECT JSON_MODIFY(@garage, '$.Cars[0]', NULL)
-- Output: { "Cars": [null, { "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" },{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GTI" }, "Year": 2017, "PurchaseDate": "2017-02-19T00:00:00.000Z" }] }

-- To truly delete it (and not have the NULL appear as the first item in the array) we have to convert to a rowset, select everything that's not the first row, aggregate the rows into a string (UGH) and then recreate as JSON.
-- This is incredibly ugly.  The STREAM_AGG() function in SQL vNext should make it a little cleaner, but why doesn't the JSON_MODIFY NULL syntax just get rid of the element in the array?
-- I have opened a Microsoft connect issue for this here: https://connect.microsoft.com/SQLServer/feedback/details/3120404 
SELECT JSON_QUERY('{ "Cars" : [' + 
        STUFF((
               SELECT   ',' + value
               FROM OPENJSON(@garage, '$.Cars') 
               WHERE [key] <> 0
               FOR XML PATH('')), 1, 1, '') + '] }')
-- Output: { "Cars" : [{ "Make": "Subaru", "Model": { "Base": "Impreza", "Trim": "Premium" }, "Year": 2016, "PurchaseDate": "2015-08-18T00:00:00.000Z" },{ "Make": "Volkswagen", "Model": { "Base": "Golf", "Trim": "GTI" }, "Year": 2017, "PurchaseDate": "2017-02-19T00:00:00.000Z" }] }

SQL JSON Performance Tuning

SQL JSON functions are already fast. Adding computed columns and indexes makes them extremely fast.

Computed Column JSON Indexes

JSON indexes are simply regular indexes on computed columns.

Add a computed column:

-- Car data source: https://github.com/arthurkao/vehicle-make-model-data
IF OBJECT_ID('dbo.Cars') IS NOT NULL 
BEGIN
    DROP TABLE dbo.Cars;
END
CREATE TABLE dbo.Cars
(
    Id INT IDENTITY(1,1),
    CarDetails NVARCHAR(MAX)
);
-- See https://gist.github.com/bertwagner/1df2531676112c24cd1ab298fc750eb2 for the full untruncated version of this code
DECLARE @cars nvarchar(max) = '[ {"year":2001,"make":"ACURA","model":"CL"}, {"year":2001,"make":"ACURA","model":"EL"},...]';

INSERT INTO dbo.Cars (CarDetails)
SELECT value FROM OPENJSON(@cars, '$');

SELECT * FROM dbo.Cars;
/* 
Output:
Id          CarDetails
----------- ----------------------------------------------
1           {"year":2001,"make":"ACURA","model":"CL"}
2           {"year":2001,"make":"ACURA","model":"EL"}
3           {"year":2001,"make":"ACURA","model":"INTEGRA"}
...
*/

-- Remember to turn on "Include Actual Execution Plan" for all of these examples

-- Before we add any computed columns/indexes, let's see our execution plan for our SQL statement with a JSON predicate
SELECT * FROM dbo.Cars WHERE JSON_VALUE(CarDetails, '$.model') = 'Golf'
/*
Output:
Id          CarDetails
----------- --------------------------------------------------
1113        {"year":2001,"make":"VOLKSWAGEN","model":"GOLF"}
2410        {"year":2002,"make":"VOLKSWAGEN","model":"GOLF"}
3707        {"year":2003,"make":"VOLKSWAGEN","model":"GOLF"}
...
*/
-- The execution plan shows a Table Scan, not very efficient

-- We can now add a non-persisted computed column for our "model" JSON property.
ALTER TABLE dbo.Cars
ADD CarModel AS JSON_VALUE(CarDetails, '$.model');

-- We add the distinct to avoid parameter sniffing issues.  
-- Our execution plan now shows the extra computation that is occuring for every row of the table scan.
SELECT DISTINCT * FROM dbo.Cars WHERE JSON_VALUE(CarDetails, '$.model') = 'Golf'
SELECT DISTINCT * FROM dbo.Cars WHERE CarModel = 'Golf'

Add an index to our computed column:

-- Add an index onto our computed column
CREATE CLUSTERED INDEX CL_CarModel ON dbo.Cars (CarModel)

-- Check the execution plans again
SELECT DISTINCT * FROM dbo.Cars WHERE JSON_VALUE(CarDetails, '$.model') = 'Golf'
SELECT DISTINCT * FROM dbo.Cars WHERE CarModel = 'Golf'
-- We now get index seeks!

-- Indexed computed column returns results in ~1ms
SELECT * FROM dbo.Cars WHERE CarModel = 'Golf'

Performance test:

-- Turn on stats and see how long it takes to parse the ~20k JSON array elements
SET STATISTICS TIME ON

-- Test #1
-- Test how long it takes to parse each property from all ~20k elements from the JSON array
-- SQL returns this query in ~546ms
SELECT JSON_VALUE(value, '$.year') AS [Year], JSON_VALUE(value, '$.make') AS Make, JSON_VALUE(value, '$.model') AS Model FROM OPENJSON(@cars, '$') 

-- Test #2
-- Time to deserialize and query just Golfs without computed column + index
-- This takes ~255ms in SQL Server
SELECT * FROM OPENJSON(@cars, '$') WHERE JSON_VALUE(value, '$.model') = 'Golf'

-- Test #3
-- Time it takes to compute the same query for Golf's with a computed column and clustered index 
-- This takes ~1ms on SQL Server
SELECT * FROM dbo.Cars WHERE CarModel = 'Golf'

-- Test #4
-- Serializing data on SQL Server takes ~110ms
SELECT * FROM dbo.Cars FOR JSON AUTO

-- What about serializing/deserializing smaller JSON datasets?
-- Let's create our smaller set
DECLARE @carsSmall nvarchar(max) = '[ {"year":2001,"make":"ACURA","model":"CL"}, {"year":2001,"make":"ACURA","model":"EL"}, {"year":2001,"make":"ACURA","model":"INTEGRA"}, {"year":2001,"make":"ACURA","model":"MDX"}, {"year":2001,"make":"ACURA","model":"NSX"}, {"year":2001,"make":"ACURA","model":"RL"}, {"year":2001,"make":"ACURA","model":"TL"}]';

-- Test #5
-- Running our query results in the data becoming deserialized in ~0ms
SELECT JSON_VALUE(value, '$.year') AS [Year], JSON_VALUE(value, '$.make') AS Make, JSON_VALUE(value, '$.model') AS Model FROM OPENJSON(@carsSmall, '$') 
--30ms in sql

-- Test #6
-- And serialized in ~0ms
SELECT TOP 7  * FROM dbo.Cars FOR JSON AUTO

JSON Support Is The Best New Developer Feature in SQL 2016 — Part 4: Performance Comparisons

Published Tue 21 February 2017 in SQL > Development > JSON

c8d32-1uzeick0fj7xna6ua0qzypg

This is the fourth article in my series about learning how to use SQL Server 2016's new JSON functions. If you haven't already, you can read Part 1 — Parsing JSON, Part 2 — Creating JSON, and Part 3 — Updating, Adding, and Deleting JSON.

Additional performance comparisons available in an updated post.

We've finally come to my favorite part of analyzing any new software feature: performance testing. SQL Server 2016's new JSON functions are great for parsing JSON data, creating JSON data, and modifying JSON data, but are they efficient?

Today we'll examine three areas of SQL Server JSON performance:

How to maximize performance for the new SQL Server JSON functions
How the new SQL Server JSON functions compare against what was previously available in SQL Server
How the new SQL JSON functions compare against Newtonsoft's Json.NET

Maximizing SQL Server JSON Function Performance

I wanted to use a sizable data set in order to test the performance of the new JSON functions in SQL Server 2016. I found arthurkao's car year/make/model data on GitHub and decided this ~20k element JSON array would be perfect for performance testing purposes. For my tests I'll be using both the original JSON string as well as a SQL table that I created from the original JSON array:

-- Car data source: https://github.com/arthurkao/vehicle-make-model-data
IF OBJECT_ID('dbo.Cars') IS NOT NULL 
BEGIN
    DROP TABLE dbo.Cars;
END
CREATE TABLE dbo.Cars
(
    Id INT IDENTITY(1,1),
    CarDetails NVARCHAR(MAX)
);
-- See https://gist.github.com/bertwagner/1df2531676112c24cd1ab298fc750eb2 for the full untruncated version of this code
DECLARE @cars nvarchar(max) = '[ {"year":2001,"make":"ACURA","model":"CL"}, {"year":2001,"make":"ACURA","model":"EL"},...]';

INSERT INTO dbo.Cars (CarDetails)
SELECT value FROM OPENJSON(@cars, '$');

SELECT * FROM dbo.Cars;
/* 
Output:
Id          CarDetails
----------- ----------------------------------------------
1           {"year":2001,"make":"ACURA","model":"CL"}
2           {"year":2001,"make":"ACURA","model":"EL"}
3           {"year":2001,"make":"ACURA","model":"INTEGRA"}
...
*/

Unlike XML in SQL Server (which is stored in it's own datatype), JSON in SQL Server 2016 is stored as an NVARCHAR. This means instead of needing to use special indexes, we can use indexes that we are already familiar with.

To maximize performance, we can use Microsoft's recommendation of adding a computed column for one of the JSON properties and then indexing that computed column:

-- Remember to turn on "Include Actual Execution Plan" for all of these examples

-- Before we add any computed columns/indexes, let's see our execution plan for our SQL statement with a JSON predicate
SELECT * FROM dbo.Cars WHERE JSON_VALUE(CarDetails, '$.model') = 'Golf'
/*
Output:
Id          CarDetails
----------- --------------------------------------------------
1113        {"year":2001,"make":"VOLKSWAGEN","model":"GOLF"}
2410        {"year":2002,"make":"VOLKSWAGEN","model":"GOLF"}
3707        {"year":2003,"make":"VOLKSWAGEN","model":"GOLF"}
...
*/
-- The execution plan shows a Table Scan, not very efficient

-- We can now add a non-persisted computed column for our "model" JSON property.
ALTER TABLE dbo.Cars
ADD CarModel AS JSON_VALUE(CarDetails, '$.model');

-- We add the distinct to avoid parameter sniffing issues.  
-- Our execution plan now shows the extra computation that is occuring for every row of the table scan.
SELECT DISTINCT * FROM dbo.Cars WHERE JSON_VALUE(CarDetails, '$.model') = 'Golf'
SELECT DISTINCT * FROM dbo.Cars WHERE CarModel = 'Golf'

Non-persisted computed columns (like in the example above) do not take up any additional space in the table. You can verify this for yourself by running sp_spaceused 'dbo.Cars' before and after adding the non-persisted column to the table.

Having a computed column doesn't add any performance to our query on its own but it does now allow us to add an index to our parsed/computed JSON property.

Having the computed column doesn't improve performance — we are still seeing a Table Scan

The clustered index that we add next stores pointers to each parsed/computed value causing the table not to take up any space and only causes the SQL engine to recompute the columns when the index needs to be rebuilt:

-- Add an index onto our computed column
CREATE CLUSTERED INDEX CL_CarModel ON dbo.Cars (CarModel)

-- Check the execution plans again
SELECT DISTINCT * FROM dbo.Cars WHERE JSON_VALUE(CarDetails, '$.model') = 'Golf'
SELECT DISTINCT * FROM dbo.Cars WHERE CarModel = 'Golf'
-- We now get index seeks!

And the resulting execution plan now shows both queries (the one using JSON_VALUE() in the WHERE clause directly as well the one calling our computed column) using index seeks to find the data we are looking for:

Yay index seeks!

Overall, adding computed columns to our table adds no overhead in terms of storage space and allows us to then add indexes on JSON properties which improve performance significantly.

SQL Server 2016 JSON vs SQL Server pre-2016 JSON

As I've mentioned before, the best option for processing JSON data in SQL Server before 2016 was by using Phil Factor's amazing JSON parsing function. Although the function works well, it is limited by what SQL Server functionality was available at the time and therefore wasn't all that efficient.

-- Let's compare how quick Phil Factor's JSON parsing function does against the new SQL 2016 functions
-- Phil's parseJSON function can be downloaded from https://www.simple-talk.com/sql/t-sql-programming/consuming-json-strings-in-sql-server/

SELECT years.StringValue AS Year, makes.StringValue AS Make, models.StringValue AS Model FROM dbo.parseJSON(@cars) models
INNER JOIN dbo.parseJSON(@cars) years ON models.parent_ID = years.parent_ID
INNER JOIN dbo.parseJSON(@cars) makes ON models.parent_ID = makes.parent_ID
WHERE models.NAME = 'model' AND models.StringValue = 'Golf' AND years.NAME = 'year' AND makes.NAME = 'make'

The above query should work for getting the data we need. I'm abusing what the parseJSON function was probably built to do (I don't think it was intended to parse ~20k element JSON arrays), and I'll be honest I waited 10 minutes before killing the query. Basically, trying to parse this much data in SQL before 2016 just wasn't possible (unless you wrote CLR).

Compared to the following queries which is using our indexed computed column SQL Server 2016 is able to return all of the results to us in 1 ms:

-- Indexed computed column returns results in ~1ms
SELECT * FROM dbo.Cars WHERE CarModel = 'Golf'

SQL Server 2016 JSON vs Newtonsoft's Json.NET

In cases like the above where parsing JSON in SQL Server was never an option, my preferred method has always been to parse data in C#. In particular, Newtonsoft's Json.NET is the standard for high performance JSON parsing, so let's take a look at how SQL Server 2016 compares to that.

The following code shows 6 tests I ran in SQL Server 2016:

-- Turn on stats and see how long it takes to parse the ~20k JSON array elements
SET STATISTICS TIME ON

-- Test #1
-- Test how long it takes to parse each property from all ~20k elements from the JSON array
-- SQL returns this query in ~546ms
SELECT JSON_VALUE(value, '$.year') AS [Year], JSON_VALUE(value, '$.make') AS Make, JSON_VALUE(value, '$.model') AS Model FROM OPENJSON(@cars, '$') 

-- Test #2
-- Time to deserialize and query just Golfs without computed column + index
-- This takes ~255ms in SQL Server
SELECT * FROM OPENJSON(@cars, '$') WHERE JSON_VALUE(value, '$.model') = 'Golf'

-- Test #3
-- Time it takes to compute the same query for Golf's with a computed column and clustered index 
-- This takes ~1ms on SQL Server
SELECT * FROM dbo.Cars WHERE CarModel = 'Golf'

-- Test #4
-- Serializing data on SQL Server takes ~110ms
SELECT * FROM dbo.Cars FOR JSON AUTO

-- What about serializing/deserializing smaller JSON datasets?
-- Let's create our smaller set
DECLARE @carsSmall nvarchar(max) = '[ {"year":2001,"make":"ACURA","model":"CL"}, {"year":2001,"make":"ACURA","model":"EL"}, {"year":2001,"make":"ACURA","model":"INTEGRA"}, {"year":2001,"make":"ACURA","model":"MDX"}, {"year":2001,"make":"ACURA","model":"NSX"}, {"year":2001,"make":"ACURA","model":"RL"}, {"year":2001,"make":"ACURA","model":"TL"}]';

-- Test #5
-- Running our query results in the data becoming deserialized in ~0ms
SELECT JSON_VALUE(value, '$.year') AS [Year], JSON_VALUE(value, '$.make') AS Make, JSON_VALUE(value, '$.model') AS Model FROM OPENJSON(@carsSmall, '$') 
--30ms in sql

-- Test #6
-- And serialized in ~0ms
SELECT TOP 7  * FROM dbo.Cars FOR JSON AUTO

And then the same tests in a C# console app using Json.Net:

static void Main(string[] args)
{
    string cars = @"[ {""year"":2001,""make"":""ACURA"",""model"":""CL""}, ... ]";
    Stopwatch stopwatch = new Stopwatch();

    // Test #1
    stopwatch.Start();
    var deserializedCars = JsonConvert.DeserializeObject<IEnumerable<Car>>(cars);
    stopwatch.Stop();
    long elapsedMillisecondsDeserialize = stopwatch.ElapsedMilliseconds;

    // Test #2 & #3
    stopwatch.Restart();
    var queriedCars = JsonConvert.DeserializeObject<IEnumerable<Car>>(cars).Where(x=>x.model == "Golf");
    stopwatch.Stop();
    long elapsedMillisecondsQuery = stopwatch.ElapsedMilliseconds;

    // Test #4
    stopwatch.Restart();
    var serializedCars = JsonConvert.SerializeObject(deserializedCars);
    stopwatch.Stop();
    long elapsedMillisecondsSerialize = stopwatch.ElapsedMilliseconds;

    // smaller data
    string carsSmall = @"[ {""year"":2001,""make"":""ACURA"",""model"":""CL""}, {""year"":2001,""make"":""ACURA"",""model"":""EL""}, {""year"":2001,""make"":""ACURA"",""model"":""INTEGRA""}, {""year"":2001,""make"":""ACURA"",""model"":""MDX""}, {""year"":2001,""make"":""ACURA"",""model"":""NSX""}, {""year"":2001,""make"":""ACURA"",""model"":""RL""}, {""year"":2001,""make"":""ACURA"",""model"":""TL""}]";

    // Test #5
    stopwatch.Restart();
    var deserializedCarsSmall = JsonConvert.DeserializeObject<IEnumerable<Car>>(carsSmall);
    stopwatch.Stop();
    long elapsedMillisecondsDeserializeSmall = stopwatch.ElapsedMilliseconds;

    // Test #6
    stopwatch.Restart();
    var serializedCarsSmall = JsonConvert.SerializeObject(deserializedCarsSmall);
    stopwatch.Stop();
    long elapsedMillisecondsSerializeSmall = stopwatch.ElapsedMilliseconds;
}

And the results compared side by side:

daef2-1tahnz5loihdxi59kxsh-za

Essentially, it seems like Json.Net beats SQL Server 2016 on larger JSON manipulations, both are equal with small JSON objects, and SQL Server 2016 has the advantage at filtering JSON data when indexes are used.

Conclusion

SQL Server 2016 is excellent at working with JSON. Even though Json.NET beats SQL Server 2016 at working with large JSON objects (on the magnitude of milliseconds), SQL Server is equally fast on smaller objects and is advantageous when JSON data needs to be filtered or searched.

I look forward to using the SQL Server 2016 JSON functions more in the future, especially in instances where network I/O benefits me to process JSON on the SQL Server or when working with applications that cannot process JSON data, like SQL Server Reporting Services.

How SQL 2016's JSON Functionality Will Make You Want To Use JSON

Published Tue 14 February 2017 in SQL > TSqlTuesday

00ad3-1lh0mvkliatliiikt0vlyow

This post is a response to this month's T-SQL Tuesday prompt. T-SQL Tuesday was created by Adam Machanic and is a way for SQL users to share ideas about interesting topics. This month's topic is Fixing Old Problems with Shiny New Toys.

Why I love JSON

I know that lots of users in the SQL community cringe when they hear about JSON, NoSQL, and other non-relational database storage options. I get it. Lots of times a string storage format compromises the many benefits SQL has to offer: indexing, data integrity, simple querying, oversight into schema changes by developers, etc… I've seen (and written) some queries that perform like molasses in winter solely due them being in a non-sargable string format.

Yet, with all of those downsides I still love JSON. I think there's a happy middle ground where DBAs and developers can agree on when it is appropriate to use JSON. Would I ever want to use JSON in a data warehouse that is being heavily queried for analytics? No. Would I also want to use JSON where performance is extremely important? No. Would I want to use it in an application where data or schema integrity is extremely important? No.

However, I love using JSON when capturing dynamically structured user input, especially if the database is being used only to persist the data instead of analyze it. I love having JSON in the database because it so easily serializes/deserializes between JSON and my C# models. If I want low latency to my webapp, I also might write the JSON from the app to the database and then have it get queued to transform into normalized data by an ETL later on. These reason in particular are what make me incredibly excited for the new JSON functionality in SQL Server 2016.

Okay, I guess there are some valid uses for JSON. What are my options for using JSON in SQL Server?

Before SQL Server 2016 was released there was no support for JSON in SQL Server. None. Zilch. Zero. The best thing we had was Phil Factor's amazing JSON parsing user defined function or writing your own CLR.

Phil's function is truly amazing and it cleverly uses some built-in SQL functionality to deserialize almost any type of JSON string you can throw at it. It does however have downsides — the function has a few edge case bugs (try adding a space before any ":" in your JSON), it does not allow for easy querying of the JSON, and the function needs to be added to any database you want to use it on.

I'm not trying to downplay Phil's function — I love it and use it in a lot of my code — but the fact of the matter is that it is a very clever workaround for something that Microsoft wasn't supporting at the time.

So what did JSON parsing look like in the pre-2016 world? Let's take a look at some example data I generated using http://www.json-generator.com/:

-- This version of that data has new lines removed to save space.
-- Please see https://gist.github.com/bertwagner/965acde93706a9a5d772509e56247a1c for the neatly formatted version of the data.

DECLARE @WebsiteJson nvarchar(max) = '{ "Users": [ { "_id": "589f14e8427b0030d59615b1", "index": 0, "guid": "4a5ecd66-8c41-4553-9c1b-7597124b46e3", "isActive": true, "balance": "$1,418.71", "picture": "http://placehold.it/32x32", "age": 26, "eyeColor": "brown", "name": "Hanson Larson", "gender": "male", "company": "KENEGY", "email": "hansonlarson@kenegy.com", "phone": "+1 (804) 447-3852", "address": "698 Durland Place, Hachita, Louisiana, 1920", "about": "Consectetur laboris sunt proident ullamco ex excepteur duis cillum sit dolor occaecat officia. Sunt amet cupidatat enim mollit esse non minim dolore ullamco minim duis do. Sunt sint aliqua sit excepteur anim proident consequat magna reprehenderit laborum. Voluptate officia et duis sit laborum.\r\n", "registered": "2016-11-09T10:59:38 +05:00", "latitude": -72.754583, "longitude": 15.975401, "tags": [ "culpa", "excepteur", "adipisicing", "reprehenderit", "eiusmod", "officia", "incididunt" ], "friends": [ { "id": 0, "name": "Blake Dodson" }, { "id": 1, "name": "Marta Bullock" }, { "id": 2, "name": "Benson Soto" } ], "greeting": "Hello, Hanson Larson! You have 7 unread messages.", "favoriteFruit": "apple" }, { "_id": "589f14e8547a192d7a11a174", "index": 1, "guid": "a3c6533b-4c50-4054-9517-a03e8aa9c9ec", "isActive": true, "balance": "$3,450.53", "picture": "http://placehold.it/32x32", "age": 28, "eyeColor": "green", "name": "Wood Douglas", "gender": "male", "company": "FARMAGE", "email": "wooddouglas@farmage.com", "phone": "+1 (819) 515-3319", "address": "113 Dean Street, Kylertown, Massachusetts, 1666", "about": "Sunt laborum duis pariatur qui esse velit dolore mollit magna aute. Eiusmod velit dolor nostrud dolor culpa labore duis eiusmod sunt Lorem. Adipisicing eu minim voluptate veniam quis eu laboris dolore ipsum. Voluptate deserunt velit qui aliqua cillum. Voluptate in sit mollit irure velit fugiat dolore minim commodo. Commodo ea aliquip cupidatat commodo consequat ipsum laboris. Excepteur velit voluptate laborum dolor officia ullamco.\r\n", "registered": "2015-09-22T11:15:47 +04:00", "latitude": -88.972896, "longitude": 35.988156, "tags": [ "tempor", "in", "adipisicing", "nostrud", "officia", "cupidatat", "occaecat" ], "friends": [ { "id": 0, "name": "Vivian Gates" }, { "id": 1, "name": "Quinn Fitzgerald" }, { "id": 2, "name": "Jenifer Lowery" } ], "greeting": "Hello, Wood Douglas! You have 2 unread messages.", "favoriteFruit": "banana" }, { "_id": "589f14e8e9c6c96cd1f0b548", "index": 2, "guid": "4c71a47a-8212-46aa-97e3-bf8c136c49f5", "isActive": true, "balance": "$2,118.35", "picture": "http://placehold.it/32x32", "age": 27, "eyeColor": "green", "name": "Copeland Wiley", "gender": "male", "company": "OPTYK", "email": "copelandwiley@optyk.com", "phone": "+1 (801) 416-3699", "address": "824 Bristol Street, Sparkill, Delaware, 1706", "about": "In veniam elit nisi reprehenderit adipisicing veniam magna. Veniam proident commodo deserunt minim eiusmod commodo aliquip voluptate in occaecat id cillum fugiat. Excepteur magna adipisicing officia qui sit dolor nisi pariatur anim excepteur.\r\n", "registered": "2016-09-27T03:54:01 +04:00", "latitude": -52.752953, "longitude": -119.42916, "tags": [ "excepteur", "labore", "irure", "voluptate", "enim", "commodo", "sint" ], "friends": [ { "id": 0, "name": "Harding Sanford" }, { "id": 1, "name": "Tessa Colon" }, { "id": 2, "name": "Washington Jordan" } ], "greeting": "Hello, Copeland Wiley! You have 7 unread messages.", "favoriteFruit": "strawberry" }, { "_id": "589f14e857f47268893121d0", "index": 3, "guid": "1ba95a0c-7559-4fdd-b52a-94f5d2c76ca0", "isActive": true, "balance": "$1,955.54", "picture": "http://placehold.it/32x32", "age": 20, "eyeColor": "green", "name": "Kathleen Bradley", "gender": "female", "company": "PASTURIA", "email": "kathleenbradley@pasturia.com", "phone": "+1 (916) 463-3527", "address": "512 Holmes Lane, Chaparrito, District Of Columbia, 1651", "about": "Ea nostrud minim exercitation commodo laborum. Minim irure est sint adipisicing. Voluptate nisi fugiat qui et quis sint nisi ea deserunt laboris est dolor cillum. Ad sit duis nisi culpa aliquip incididunt.\r\n", "registered": "2015-07-20T03:11:35 +04:00", "latitude": -9.512905, "longitude": -108.671577, "tags": [ "eiusmod", "enim", "tempor", "amet", "reprehenderit", "sit", "ut" ], "friends": [ { "id": 0, "name": "Rowe Crawford" }, { "id": 1, "name": "Ginger Bray" }, { "id": 2, "name": "Nichole Bradford" } ], "greeting": "Hello, Kathleen Bradley! You have 6 unread messages.", "favoriteFruit": "banana" }, { "_id": "589f14e8128844a31ccb82e2", "index": 4, "guid": "03e2341e-da4d-48bb-8d6f-006234835828", "isActive": false, "balance": "$1,827.38", "picture": "http://placehold.it/32x32", "age": 32, "eyeColor": "brown", "name": "Colon Mays", "gender": "male", "company": "FREAKIN", "email": "colonmays@freakin.com", "phone": "+1 (919) 555-3410", "address": "627 Reed Street, Taft, Texas, 1724", "about": "Ullamco magna tempor dolore enim veniam aute incididunt. Deserunt nostrud fugiat reprehenderit consequat deserunt. Labore consequat magna pariatur amet mollit aute reprehenderit ea dolore exercitation anim.\r\n", "registered": "2014-06-08T02:13:34 +04:00", "latitude": 35.186997, "longitude": -1.722535, "tags": [ "cillum", "amet", "ad", "enim", "tempor", "amet", "est" ], "friends": [ { "id": 0, "name": "Gladys Rojas" }, { "id": 1, "name": "Luann Craft" }, { "id": 2, "name": "Camille Gibson" } ], "greeting": "Hello, Colon Mays! You have 7 unread messages.", "favoriteFruit": "apple" } ] }';

Basically we have a JSON string that represents five users and some of their attributes (note: I would argue that storing this much data about an individual would be better suited to fit into some structured tables, but alas this was the type of data that was easy to generate for example purposes).

If I wanted to parse out each user's name to use for an infrequently ran report or for ETL purposes, this is what my query would have to look like:

SELECT 
    NAME, 
    StringValue, 
    ValueType
FROM 
(
    SELECT 
        [NAME], 
        StringValue, 
        ValueType, 
        LAG([NAME],1,0) OVER (ORDER BY  element_id, sequenceNo, parent_ID) as PreviousKey 
    FROM 
        Sandbox.dbo.parseJSON(@WebsiteJson) -- This is Phil Factor's JSON parsing function: https://www.simple-talk.com/sql/t-sql-programming/consuming-json-strings-in-sql-server/
)d 
WHERE 
    PreviousKey <> 'id' 
    and [NAME] = 'name'

There are a couple funny things we have to do in pre-2016 SQL to parse this JSON:

We have to add the parseJson() function to our database
Since there are multiple key's named "NAME", we need to add a LAG() function to help identify the property's key so we can ensure we have the correct one.
Nested queries required since windowed functions can't run in the WHERE clause.

Overall this code works and it gets the job done. However we need to ensure our server has the parseJson( function added and filtering our JSON data to pull out the properties we are interested in takes a little bit of work. Additionally, the parseJson() function does do some heavy processing which affects our overall performance:

25f05-1bljeawm2hatdamfym6na3w

The Shiny New Toys in SQL Server 2016: JSON Functions

SQL Server 2016 includes many new JSON parsing functions that are available for us to use. Writing a query to return the same result set as above is now as simple as this:

SELECT 
    JSON_VALUE(value, '$.name') 
FROM 
    OPENJSON(@WebsiteJson,'$.Users')

The OPENJSON() function allows us to write an XPath query to filter the JSON Users object into five separate rows of data, one per user:

445e4-1l0a1zh0rvyy6x06goaiwhq

Then, we use the JSON_VALUE() function and XPath once again to filter out just the "name" property. Overall, this code is much simpler to write, performs more consistently (the issue with a space character before a ":" is handled correctly), and performs much quicker:

17673-15sbboqewqoahnr8vayguqg

I love the syntax of the new SQL JSON parsing functions: it's easy to remember, easy to use in real life scenarios, and is very fast. This is one new feature of SQL Server 2016 that definitely makes my life easier and makes any other ways of parsing out JSON data obsolete.

If you are interested in learning more about JSON in SQL server, I've been blogging the past few weeks about how to use all of SQL's new JSON functions.

There's also one bug that I found with how SQL handles JSON deletes that I submitted to Microsoft Connect if you want to upvote and see JSON continue to evolve on SQL Server.