topic Re: globalMap.get() is not working in Talend Studio

globalMap.get() is not working

TomG1 — Sat, 16 Nov 2024 09:41:20 GMT

I am using a global map variable in a tJavaRow component. but its giving an error message.

"cannot make a static reference to the non-static field".

I am using BigData Spark job.

How Do i overcome this issue.

Re: globalMap.get() is not working

cterenzi — Fri, 02 Jun 2017 19:04:59 GMT

Can you show your component code instead of the Code tab?

Re: globalMap.get() is not working

TomG1 — Sun, 04 Jun 2017 12:35:57 GMT

I have created a simple mapping to show this error.

the error is coming from the tJavaRow component which i have shown in the first post.

here is the code of the tjavarow component .

 String[] values = (String[])globalMap.get("values");
 
 row1.a = row2.a;
 row1.b = row2.b;

here is the code in tjava component

String[] values = { "This", "is", "test" };

globalMap.put("values", values);

try reproducing the error in you system.

request to help me on this..

Re: globalMap.get() is not working

Anonymous — Sun, 04 Jun 2017 17:43:37 GMT

@tomaszc, I tried the way you explained. For me it is working fine, i was able to get the values from globalMap in the tJavaRow.

I've exported my job (Test.zip), you can download it and have a look.

Hope this helps.

Re: globalMap.get() is not working

TomG1 — Sun, 04 Jun 2017 19:37:57 GMT

Hi,

I am not able to open your job in talend.

what kind of job is that.. is it standard or bigdata job?.

I am specifically talking about Bigdata Spark job.

In Talend standard job , I am able to get the output . but the same code in talend big data is giving error.

I am using 'Talend big data plaform 6.2'

Thanks

Re: globalMap.get() is not working

Anonymous — Mon, 05 Jun 2017 02:01:06 GMT

Sorry it was a standard job, not a Bigdata Spark job. I will give it a try with Bigdata Spark job.

Re: globalMap.get() is not working

TomG1 — Mon, 05 Jun 2017 08:02:34 GMT

I am trying with an alternative

I have created a context variable of type Object.

To this variable I am saving my string array in tJava component

String[] cols = { "This","is","test" };
context.SrcCols = cols;

I am accessing this context variable in tJavarow

String[] Cols = (String[])context.SrcCols;

but while running its giving the following error.

java.lang.ClassCastException: java.lang.String cannot be cast to [Ljava.lang.String;

Re: globalMap.get() is not working

Anil3 — Tue, 19 Dec 2017 16:21:29 GMT

Hello Tom,

Did you get any solution to make it work in BigData Spark Job? we tried to implement the same and found this artilce with the same error what we are getting.

Please suggest.

Re: globalMap.get() is not working

Anonymous — Tue, 19 Dec 2017 17:45:12 GMT

Hello,

Currently, there is no access for "globalMap" in Spark Batch mode. This is due to a different implementation of Spark Batch compared to DI. You will only be able to accomplish this with standard DI jobs.

Re: globalMap.get() is not working

TomG1 — Mon, 05 Feb 2018 09:57:22 GMT

This feature is not available in talend spark jobs.

We have raised a ticket to talend support and they mentioned that it is a limitation in talend big data job.

I don't know if this feature is available in talend latest version

Thanks

Re: globalMap.get() is not working

Anonymous — Sat, 14 Apr 2018 23:07:45 GMT

I need a globalmap or an alternative to this in spark jobs. Any ideas?

Re: globalMap.get() is not working

Anil3 — Sun, 15 Apr 2018 18:52:12 GMT

Hello Guys, I opened a ticket with Talend and Jack from talend already responded to this thread.. So we went and tried something different

--> Create context variable solution like TOM suggested.

--> Instead of usting tJavaRow, use tJava then you will not get the casting error.

--> your job is somewaht like source data component --> Row (Main) --> tJava

--> I know tJava we should not use for main flow however Spark dataframe processing is different than actual main row flow in other talend jobs.

--> with this solution we are able to make Spark job work but we couldn't take this solution to production as we are not convinced to use tJava like this.