einvoice | ระบบพิมพ์ใบแจ้งหนี้/ใบเสร็จรับเงิน/ใบกำกับภาษี

In your json file you first have the json object called errors, which has the array errors. But your response class contains the property of type List, so I think it should have a property of a type which would define a property of type List.

Change your errors class to something like this and add a new class to represent a single error. I think it should work.

async and await have been introduced in C# 5. Visual Studio 2010 supports C# 4. Thats why you can't compile code with async/await under VS 2010.

To suggest a switch to HttpClient if possible:

I left out the header-configs for brevity, but it's definitely possible to do.

Please list the jars you added.

See here: https://www.cwiki.us/questions/13469089/could-not-initialize-class-org.apache.logging.log4j.logmanager

I would use jsonconverter.bas to parse the json. After installing the code from that link in a standard module called JsonConverter, go to VBE > Tools > References > Add a reference to Microsoft Scripting Runtime.

Then I would dimension an array to hold the results. I would determine rows from the number of items in the json collection returned and the number of columns from the size of the first item dictionary. Loop the json object, and inner loop the dictionary keys of each dictionary in collection, and populate the array. Write the array out in one go at end.

Below, I am reading in the json string from cell A1 but you would replace that with your json source.

I am not familiar with that format so I can't help with deciphering your deadlock. It looks like a text version of a deadlock graph.

There is always going to be more than one statement involved in a deadlock. In the sql server world a deadlock is a competition for resources(locks) between two or more processes. Almost every time a deadlock victim is chosen, it will be a select statement. Select commands (reads) hold the least amount of resources and are the easiest to rollback. Sql server will rollback the process holding the least amount of resources. This is probably why your reporting procedure was the chosen victim. The commands from other processes you are seeing were all active queries when you ran the report.

Reading a deadlock graph is not trivial. Deadlocks usually occur when long read processes are running during a time of heavy writes. There are ways to attempt to prevent deadlocks. If you do a web search for deadlock prevention then you will find some useful information on the subject, however, No matter what precautions you take, you are going to run into one or many deadlocks when your system gets busy enough.

Ideally, you would try and split up a write heavy and read heavy system into two worlds, OLTP and OLAP, however, this is not always practical.

If your report will not be adversely affected by possibly reading uncommitted data then you can avoid this problem in your reporting stored procedures by placing NOLOCK hints on your joins and table select statements or specifying a read uncommitted transaction isolation level for your stored procedure. This is not always ideal but neither are deadlocks :)

a4j:commandLink has ajax already. So use execute from a4j:commandLink directly. Don't use another ajax tag. You can't (or, for sure, you shouldn't) use both.

Or I you really want to use f:ajax, then use h:commandLink instead of a4j:commandLink.

PS. You can try to use a4j:ajax, but I'm pretty sure the result will be the same as for f:ajax.