Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

Hi Adam,

As you mentioned that when you run the Flow Service the variable is not available post service execution indicating the mapping is not working.

Have you tried debugging the flow service and immediately after the Invoke step where you have performed the mapping do you see the mapped variable ? Any Dropping of variables in the subsequent steps post the mapping? Is pub.flow:clearPipeline service being invoked in the flow service and hence your variable is cleared from the pipeline?

Perhaps if you can share Screenshots of the mapping and/or the Flow service it might help to assist

------------------------------
Kailash Kumar Mishra

Original Message:
Sent: Tue March 04, 2025 03:19 PM
From: Adam Connor
Subject: Understanding variables in pipelines

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

The service is simple:

The mapping step takes a single field, LAST-ISN, and tries to map it to "last_isn":

If I run it in the debugger, after MAP I see

(no "last_isn") which is the same thing I see if I run it normally and look at the results.

Hi Adam,

As you mentioned that when you run the Flow Service the variable is not available post service execution indicating the mapping is not working.

Have you tried debugging the flow service and immediately after the Invoke step where you have performed the mapping do you see the mapped variable ? Any Dropping of variables in the subsequent steps post the mapping? Is pub.flow:clearPipeline service being invoked in the flow service and hence your variable is cleared from the pipeline?

Perhaps if you can share Screenshots of the mapping and/or the Flow service it might help to assist

------------------------------
Kailash Kumar Mishra

Original Message:
Sent: Tue March 04, 2025 03:19 PM
From: Adam Connor
Subject: Understanding variables in pipelines

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

Hi Adam,

Your service is indeed simple and the mapping should work ideally. 

Have noticed that the Source Field LAST_IN is of number type ( has 64 mentioned ) whereas Target Variable last_in is normal string ( has Abc mentioned ) not sure if this is preventing the mapping from taking place.

Perhaps you can make the variable last_in too as numeric type 64  of the same type as Source field and try once and see if the mapping works

The service is simple:

The mapping step takes a single field, LAST-ISN, and tries to map it to "last_isn":

If I run it in the debugger, after MAP I see

(no "last_isn") which is the same thing I see if I run it normally and look at the results.

Hi Adam,

As you mentioned that when you run the Flow Service the variable is not available post service execution indicating the mapping is not working.

Have you tried debugging the flow service and immediately after the Invoke step where you have performed the mapping do you see the mapped variable ? Any Dropping of variables in the subsequent steps post the mapping? Is pub.flow:clearPipeline service being invoked in the flow service and hence your variable is cleared from the pipeline?

Perhaps if you can share Screenshots of the mapping and/or the Flow service it might help to assist

------------------------------
Kailash Kumar Mishra

Original Message:
Sent: Tue March 04, 2025 03:19 PM
From: Adam Connor
Subject: Understanding variables in pipelines

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

In truth, those indicators are too small to really see on a high resolution monitor. Interesting. When you add a variable to a pipeline, you can't really specify a type like Int64. I can make is an "object" type, though. And then it works, so thank you. Still disturbing that there is no error raised.

Hi Adam,

Your service is indeed simple and the mapping should work ideally. 

Have noticed that the Source Field LAST_IN is of number type ( has 64 mentioned ) whereas Target Variable last_in is normal string ( has Abc mentioned ) not sure if this is preventing the mapping from taking place.

Perhaps you can make the variable last_in too as numeric type 64  of the same type as Source field and try once and see if the mapping works

The service is simple:

The mapping step takes a single field, LAST-ISN, and tries to map it to "last_isn":

If I run it in the debugger, after MAP I see

(no "last_isn") which is the same thing I see if I run it normally and look at the results.

Hi Adam,

As you mentioned that when you run the Flow Service the variable is not available post service execution indicating the mapping is not working.

Have you tried debugging the flow service and immediately after the Invoke step where you have performed the mapping do you see the mapped variable ? Any Dropping of variables in the subsequent steps post the mapping? Is pub.flow:clearPipeline service being invoked in the flow service and hence your variable is cleared from the pipeline?

Perhaps if you can share Screenshots of the mapping and/or the Flow service it might help to assist

------------------------------
Kailash Kumar Mishra

Original Message:
Sent: Tue March 04, 2025 03:19 PM
From: Adam Connor
Subject: Understanding variables in pipelines

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

In truth, those indicators are too small to really see on a high resolution monitor. Interesting. When you add a variable to a pipeline, you can't really specify a type like Int64. I can make is an "object" type, though. And then it works, so thank you. Still disturbing that there is no error raised.

Hi Adam,

Your service is indeed simple and the mapping should work ideally. 

Have noticed that the Source Field LAST_IN is of number type ( has 64 mentioned ) whereas Target Variable last_in is normal string ( has Abc mentioned ) not sure if this is preventing the mapping from taking place.

Perhaps you can make the variable last_in too as numeric type 64  of the same type as Source field and try once and see if the mapping works

The service is simple:

The mapping step takes a single field, LAST-ISN, and tries to map it to "last_isn":

If I run it in the debugger, after MAP I see

(no "last_isn") which is the same thing I see if I run it normally and look at the results.

Hi Adam,

As you mentioned that when you run the Flow Service the variable is not available post service execution indicating the mapping is not working.

Have you tried debugging the flow service and immediately after the Invoke step where you have performed the mapping do you see the mapped variable ? Any Dropping of variables in the subsequent steps post the mapping? Is pub.flow:clearPipeline service being invoked in the flow service and hence your variable is cleared from the pipeline?

Perhaps if you can share Screenshots of the mapping and/or the Flow service it might help to assist

------------------------------
Kailash Kumar Mishra

Original Message:
Sent: Tue March 04, 2025 03:19 PM
From: Adam Connor
Subject: Understanding variables in pipelines

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

The service is simple:

The mapping step takes a single field, LAST-ISN, and tries to map it to "last_isn":

If I run it in the debugger, after MAP I see

(no "last_isn") which is the same thing I see if I run it normally and look at the results.

Hi Adam,

As you mentioned that when you run the Flow Service the variable is not available post service execution indicating the mapping is not working.

Have you tried debugging the flow service and immediately after the Invoke step where you have performed the mapping do you see the mapped variable ? Any Dropping of variables in the subsequent steps post the mapping? Is pub.flow:clearPipeline service being invoked in the flow service and hence your variable is cleared from the pipeline?

Perhaps if you can share Screenshots of the mapping and/or the Flow service it might help to assist

------------------------------
Kailash Kumar Mishra

Original Message:
Sent: Tue March 04, 2025 03:19 PM
From: Adam Connor
Subject: Understanding variables in pipelines

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

I'm thinking it's because you are mapping from a double to a string. I suggest adding an objectToString transformer to your map.

The service is simple:

The mapping step takes a single field, LAST-ISN, and tries to map it to "last_isn":

If I run it in the debugger, after MAP I see

(no "last_isn") which is the same thing I see if I run it normally and look at the results.

Hi Adam,

As you mentioned that when you run the Flow Service the variable is not available post service execution indicating the mapping is not working.

Have you tried debugging the flow service and immediately after the Invoke step where you have performed the mapping do you see the mapped variable ? Any Dropping of variables in the subsequent steps post the mapping? Is pub.flow:clearPipeline service being invoked in the flow service and hence your variable is cleared from the pipeline?

Perhaps if you can share Screenshots of the mapping and/or the Flow service it might help to assist

------------------------------
Kailash Kumar Mishra

Original Message:
Sent: Tue March 04, 2025 03:19 PM
From: Adam Connor
Subject: Understanding variables in pipelines

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

Actually, at least from the UI, it isn't easy to tell what the type is. If I select the LAST-ISN field coming from Natural, I see properties that look like this:

That part is all based on the IDL parsing, but the actual field is P10 or N10, so I would think some kind of integer would be more likely.

I'll go re-read the indicated docs. I don't doubt that it's doable, although I have some reservations about whether writing flows in designer is productive.

I'm thinking it's because you are mapping from a double to a string. I suggest adding an objectToString transformer to your map.

The service is simple:

The mapping step takes a single field, LAST-ISN, and tries to map it to "last_isn":

If I run it in the debugger, after MAP I see

(no "last_isn") which is the same thing I see if I run it normally and look at the results.

Hi Adam,

As you mentioned that when you run the Flow Service the variable is not available post service execution indicating the mapping is not working.

Have you tried debugging the flow service and immediately after the Invoke step where you have performed the mapping do you see the mapped variable ? Any Dropping of variables in the subsequent steps post the mapping? Is pub.flow:clearPipeline service being invoked in the flow service and hence your variable is cleared from the pipeline?

Perhaps if you can share Screenshots of the mapping and/or the Flow service it might help to assist

------------------------------
Kailash Kumar Mishra

Original Message:
Sent: Tue March 04, 2025 03:19 PM
From: Adam Connor
Subject: Understanding variables in pipelines

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

FWIW, I can see all the types in the relevant Adapter Service. The issue is that all the data returned by the invoke, represented by the "outRec" document, does not specify those types -- it's more general. I suppose that is because the interface representing the invoke cannot know what is getting returned. But the result is that the "strong typing" of the design is kind of lost when you are mapping results from an invocation, and apparently you have to re-apply your knowledge. This seems very kludgey, because all that information was available in the adapter service in the first place.

But I will go back and re-read.

Actually, at least from the UI, it isn't easy to tell what the type is. If I select the LAST-ISN field coming from Natural, I see properties that look like this:

That part is all based on the IDL parsing, but the actual field is P10 or N10, so I would think some kind of integer would be more likely.

I'll go re-read the indicated docs. I don't doubt that it's doable, although I have some reservations about whether writing flows in designer is productive.

I'm thinking it's because you are mapping from a double to a string. I suggest adding an objectToString transformer to your map.

The service is simple:

The mapping step takes a single field, LAST-ISN, and tries to map it to "last_isn":

If I run it in the debugger, after MAP I see

(no "last_isn") which is the same thing I see if I run it normally and look at the results.

Hi Adam,

As you mentioned that when you run the Flow Service the variable is not available post service execution indicating the mapping is not working.

Have you tried debugging the flow service and immediately after the Invoke step where you have performed the mapping do you see the mapped variable ? Any Dropping of variables in the subsequent steps post the mapping? Is pub.flow:clearPipeline service being invoked in the flow service and hence your variable is cleared from the pipeline?

Perhaps if you can share Screenshots of the mapping and/or the Flow service it might help to assist

------------------------------
Kailash Kumar Mishra

Original Message:
Sent: Tue March 04, 2025 03:19 PM
From: Adam Connor
Subject: Understanding variables in pipelines

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

Hi @Adam Connor,

As of the size of the pipeline fields' annotation, yes they are indeed quite small, but believe me, once get the hang of it, you'll be able to identify them at a glance, whatever your screen resolution.

Regarding field type, yes you can specify the required Java type: once you create a new Object field (Data type: Object), from the Java wrapper type dropdown list select one of the supported Java Type, so in your case java.lang.Long

FWIW, I can see all the types in the relevant Adapter Service. The issue is that all the data returned by the invoke, represented by the "outRec" document, does not specify those types -- it's more general. I suppose that is because the interface representing the invoke cannot know what is getting returned. But the result is that the "strong typing" of the design is kind of lost when you are mapping results from an invocation, and apparently you have to re-apply your knowledge. This seems very kludgey, because all that information was available in the adapter service in the first place.

But I will go back and re-read.

Actually, at least from the UI, it isn't easy to tell what the type is. If I select the LAST-ISN field coming from Natural, I see properties that look like this:

That part is all based on the IDL parsing, but the actual field is P10 or N10, so I would think some kind of integer would be more likely.

I'll go re-read the indicated docs. I don't doubt that it's doable, although I have some reservations about whether writing flows in designer is productive.

I'm thinking it's because you are mapping from a double to a string. I suggest adding an objectToString transformer to your map.

The service is simple:

The mapping step takes a single field, LAST-ISN, and tries to map it to "last_isn":

If I run it in the debugger, after MAP I see

(no "last_isn") which is the same thing I see if I run it normally and look at the results.

Hi Adam,

As you mentioned that when you run the Flow Service the variable is not available post service execution indicating the mapping is not working.

Have you tried debugging the flow service and immediately after the Invoke step where you have performed the mapping do you see the mapped variable ? Any Dropping of variables in the subsequent steps post the mapping? Is pub.flow:clearPipeline service being invoked in the flow service and hence your variable is cleared from the pipeline?

Perhaps if you can share Screenshots of the mapping and/or the Flow service it might help to assist

------------------------------
Kailash Kumar Mishra

Original Message:
Sent: Tue March 04, 2025 03:19 PM
From: Adam Connor
Subject: Understanding variables in pipelines

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

Thanks, this is helpful. And I've realized I can also type the fields in documents.

TBH, what I think this part of webmethods needs is a better overview, building the mental model necessary to use webmethods successfully. It presumes a certain outlook (very Java-y, and I was a Java developer for some years, but we mostly use more dynamic languages now.)

Hi @Adam Connor,

As of the size of the pipeline fields' annotation, yes they are indeed quite small, but believe me, once get the hang of it, you'll be able to identify them at a glance, whatever your screen resolution.

Regarding field type, yes you can specify the required Java type: once you create a new Object field (Data type: Object), from the Java wrapper type dropdown list select one of the supported Java Type, so in your case java.lang.Long

FWIW, I can see all the types in the relevant Adapter Service. The issue is that all the data returned by the invoke, represented by the "outRec" document, does not specify those types -- it's more general. I suppose that is because the interface representing the invoke cannot know what is getting returned. But the result is that the "strong typing" of the design is kind of lost when you are mapping results from an invocation, and apparently you have to re-apply your knowledge. This seems very kludgey, because all that information was available in the adapter service in the first place.

But I will go back and re-read.

Actually, at least from the UI, it isn't easy to tell what the type is. If I select the LAST-ISN field coming from Natural, I see properties that look like this:

That part is all based on the IDL parsing, but the actual field is P10 or N10, so I would think some kind of integer would be more likely.

I'll go re-read the indicated docs. I don't doubt that it's doable, although I have some reservations about whether writing flows in designer is productive.

I'm thinking it's because you are mapping from a double to a string. I suggest adding an objectToString transformer to your map.

The service is simple:

The mapping step takes a single field, LAST-ISN, and tries to map it to "last_isn":

If I run it in the debugger, after MAP I see

(no "last_isn") which is the same thing I see if I run it normally and look at the results.

Hi Adam,

As you mentioned that when you run the Flow Service the variable is not available post service execution indicating the mapping is not working.

Have you tried debugging the flow service and immediately after the Invoke step where you have performed the mapping do you see the mapped variable ? Any Dropping of variables in the subsequent steps post the mapping? Is pub.flow:clearPipeline service being invoked in the flow service and hence your variable is cleared from the pipeline?

Perhaps if you can share Screenshots of the mapping and/or the Flow service it might help to assist

------------------------------
Kailash Kumar Mishra

Original Message:
Sent: Tue March 04, 2025 03:19 PM
From: Adam Connor
Subject: Understanding variables in pipelines

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

------------------------------
Adam Connor
------------------------------

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?

Here's what is confusing me, abstracted a bit. I run a simple flow with an invoke step (in this case, it invokes a Natural subprogram via RPC). It comes back with an output "Document". Great. I try to map a single field from that document (in this case, the last ISN read) to a "top-level" field. It shows the link in the designer, but when I run the flow, no such variable exists in the pipeline.

It doesn't matter either way if I make the top-level field an "output" field.

Is this possible? What am I misunderstanding about variables in pipelines?