Harald Korneliussen さんのアクティビティ

Building a complicated tree from criteria on its leaves

Harald Korneliussen — Mon, 18 Aug 2025 10:29:31 GMT

We have a product structure, fairly common I think (much of is it based on standard packages) where we have deliveries/vessels on top, one or more layers of product assemblies in between, and purchased parts (with linked manufacturer parts) at the bottom. Vessels are a separate item type, as are manufacturer parts, everything else is Part.

Now, the service department would like a view which lets them search for a specific manufacturer product - maybe one which needs replacement due to a recall, or one which requires a software update - and find all the places it's used.

The "where used" isn't good enough, because it has far too many references which aren't relevant for Service (such as all the ECOs that's ever touched the part), also it's bottom up, whereas they would like to work top-down - naturally, when servicing a part, you want to do all parts on one vessel before the next, all parts in one cabinet before the next etc.

So I'm trying to build a tree view from a search, and I find it pretty challenging. AML's recursive search with GetItemRepeatConfig doesn't work well, because if I put a condition, it demands all intermediate Parts satisfy that condition, not just the leaves. I also have to deal with a different Relationship type on the top level and on the bottom level. Finally, it requires a single item (by ID) as the root.

I imagine I'm not the only one having this sort of problem. How have you solved similar problem? Do you have a go-to method for returning (multi)trees based on conditions on their leaves?

New community edition (Release 2025)

AngelaIp — Wed, 04 Jun 2025 07:09:46 GMT

Hi community,

what are your thoughts regarding the new open release?

Angela

Multiple sessions created for trivial actions like "edit". What can be the reason?

AngelaIp — Fri, 07 Mar 2025 13:34:57 GMT

Hi community,

would be great if somebody could help me with following issue with Innovator 12SP7!

I noticed that one of my users appears with many sessions in the [application].[ACTIVEUSER] table in the database. The same user also complains that Innovator is very slow in certain scenarios. For example changing the "Make/Buy" drop-down in the Part Form can last up to 5 seconds. Opening a custom Impact Matrix with a few hundred Affected Items can take several minutes (typical would be 3-5 seconds).

I am right now confused why this specific user appears with an excess amount of sessions in the list. Typically people are only listed with a few sessions. I wonder if the amount of listed sessions and the users´ performance problems are somehow related?

The table looks like this:

Red is always the same user. Green is somebody else.

I did some tests with the user. One new connection is created when he starts Innovator. Regular searching does not create any new session. But a new session appear as soon as he opens a Parts and clicks the "Edit" button!

For each "Edit" click a new session was created! But only for this specific user. If I do administration stuff in Innovator (editing ItemTypes, Method, etc.) no additional sessions are created. I hardly appear in the session list at all.

Side notes:

- No browser extensions used. Session cookies are not blocked or removed
- We tested Firefox and Edge
- Other users only have one or a small amount of sessions
- New connections appear for file operations (thumbnails and regular file upload). But these sound legit.

Does anyone has an idea what can cause this behavior? Is this normal?

Thanks for any help!

Angela

UI doesn't load for one user - anyone had a similar problem?

Harald Korneliussen — Tue, 06 May 2025 07:45:51 GMT

On our test server, on one particular database, one particular user doesn't get an UI when logging in. I've tried digging into the frontend calls, and it seems it's this call:

https://{our test server}/Server/MetaData.asmx/GetConfigurableUi?accept=json&database={one particular db}&date=2025-05-02T13%3A01%3A58.94&item_classification=%25all_grouped_by_classification%25&item_type_id=&lang=en&location_name=TOC&user={me}

which fails, with a "not a single item" fault. I am in fact logged in, and can e.g. go see the session information on the user button. Deleting all client side caches or logging in with incognito mode makes no difference.

Have anyone else experienced this?

GetItemRepeatConfig removes child items that are used multiple times. How to fix?

AngelaIp — Fri, 20 Jan 2023 13:30:53 GMT

Hi community,

I use a classical "GetItemRepeatConfig" query to get a Part BOM over multiple levels.

id="xyz";
Item partItem = inn.newItem("Part","GetItemRepeatConfig");
partItem.setAttribute("select","id,classification,keyed_name,item_number");
partItem.setID(id); // <- target Part

Item bomItem = partItem.createRelationship("Part BOM","get");
bomItem.setAttribute("select","related_id(keyed_name),sort_order");
bomItem.setAttribute("repeatProp","related_id");
bomItem.setAttribute("repeatTimes","10");

partItem = partItem.apply();

When taking a look at the AML result, I discovered that Parts that are used on multiple levels inside a BOM (e.g. common items like screws) are not returned as complete items. The full related item is available only ONCE inside the AML result, but not for every BOM position that use that Part.

Here is how the result looks like:

The image shows three relationship BOM positions. Position 1 and 3 contain the full Part <Item> for the "related_id". Position 2 does not contain the "Item" cause it appeared already present in a previous position.

Problem: Without the date of the related_id/Item, I am not able to render the AML data. In my current use case I want to restructure the AML result in a way that it will also include depth/level of each BOM position. This basically works well, but as soon a duplicate entry appears, the data for the child item is missing:

Here is the current result that I get. I used a placeholder "Data missing" each time the related item is missing:

I discovered that this one is actually a know bug/behavior since over 10 years:

https://community.aras.com/f/archive/2097/developers-forum---getitemrepeatconfig-fix

The solution in the post is to not use GetItemRepeatConfig, but to manually create a query that includes all BOM levels.

Sure, this solution will work. But after 10 years, is this really the solution of choice? My AML query shall also contain other relationships, which would make the query even bigger when we hardcode any level manually.

Does anyone now a better solution for this one or how to modify the GetItemRepeatConfig query so it returns all data?

Thanks for any help!

Angela

What is the best way to add more lottery and gambling elements to Innovator?

AngelaIp — Tue, 01 Apr 2025 08:59:34 GMT

Hi community,

I'm looking for ideas on how to incorporate lottery, gambling or at least some game-like elements into Innovator to make certain processes more engaging. The goal is to introduce some kind of randomness, rewards, or game-like incentives within the system.

I did some research at Github and already found some promising projects like this marvelous "Chess" implementation for Innovator:

https://github.com/omkarb-sg/aras-chess

But I think there is so much more potential to showcase the flexibility of the platform! Any ideas?

TreeGridView performance

Harald Korneliussen — Thu, 27 Mar 2025 12:21:17 GMT

We have "changes" tab to list, in a tree grid view, the change objects which have gone through on a given Part.

However, loading this tab is terribly slow.

I have found out that the query itself is quite fast. When I run it in the query editor, it takes less than a second. However, the Tree Grid View can take almost 20 seconds to load. Even on a fresh part with no change items, it takes 15+ seconds.

The call which takes time is an odata call to /Server/odata/method.rb_GetTreeGridData, which takes the entire 2500+ line rb_TreeGridViewDefinition as input (why? It's a server side method, although a very odd one I can't see the implementation of, and the data it gets passed lives on the server too!)

Have anyone had any success in optimizing the loading of tree grid view structures?

Item.apply overwriting action, has it always done this?

Harald Korneliussen — Fri, 21 Feb 2025 07:40:06 GMT

After upgrading from version 12 to 31, we started getting an error when saving a certain custom datastructure, developed for us by Minerva back in the day:

Failed to execute 'serializeToString' on 'XMLSerializer': parameter 1 is not of type 'Node'.

From what I can tell, this happens because Minerva's old code calls an action directly on the document.thisItem object

var output = document.thisItem.apply("VE_GetCopyTreeGridHeader");

which "pollutes" document.thisItem with that action. When later hitting save, it calls this function instead of doing whatever it ordinarily would (probably a get action), resulting in the error. I observe that if I change the action to "get", it gets rid of the error message and saves any changes in the object itself, but not in a related items treegrid associated with it. If I change the Iaction to "update", it saves everything correctly the way it used to do in version 12.

I see that the docs on item apply says that it overwrites the action, but if it always did that, how did this code ever work? (By the way, that .apply() overwrites the action is an error-prone side effect which I can't see any good reason for)

Making changes to typescript files in the tree

Harald Korneliussen — Thu, 21 Nov 2024 11:02:29 GMT

I tried to apply a trick from a previous post:

How to open StartItem directly without the helper window - Development - Aras Community - Aras Community

However, that file is now a typescript file in my tree. I guess that means I need to transpile it to javascript/do whatever else you at Aras do in your frontend build process, in order for it to end up in the right place? What is the recommended way to do this?

ECO Item in Browser changes after voting it from Planning and to In Work

Bernt Ove Sunde — Thu, 09 Jan 2025 13:44:42 GMT

Hi all,

Currently setting up a new production environment with V31 and have gotten a very special error on our ECO item.

The ECO is slightly modified, but uses the EHT module from Minerva for dynamical assignment.

Everything is going as planned until the ECO is voted to In Work. When the ECO refreshes it has changed its keyed_name to the affected_part keyed_name, and has gotten a new item ID. When trying to seach in the database after the ECO with the new item ID it does not exists.

It seems that the ECO is now trying to be the afffected part, but the signoff tab still works, and we are able to finish the voting and release of the ECO and its affected parts.

If we close the ECO tab and open it again its back to its normal state.

It seems like its a client/browser issue, but I*m now at a miss to where troubleshoot. I have excluded server and client methods, but the issue is still present.

We are using Edge (Version 131.0.2903.112 (Official build) (64-bit))

Any tips and help is greatly appreciated!

Zombie tabs

Harald Korneliussen — Fri, 10 Jan 2025 08:08:14 GMT

At least two places in our system we experience zombie tabs, tabs that after an error message are impossible to close.

I can now reproduce the issue: We have a custom "Delivery Risks" tab on parts. If I add a new risk and open it, the UI thinks the object is in a dirty state. Just opening it and closing it again, I get a popup with "Changes have not been saved. What do you want to do?". If I select "save" here, I get the error

"The Command for the Command Alias 'SaveItem' failed."

and the form still thinks the object has unsaved changes. However, if I click "Discard", I get the following error in the web console:

Uncaught (in promise) CUIError: The Command for the Command Alias 'UnlockItem' failed.
at T0 (factory.ts:152:8)
at Object.execute (factory.ts:169:24)
at onUnlockCommand (include.aspx?classes=tearoffWindow:915:36)
at async eval (eval at x0 (factory.ts:53:9), <anonymous>:30:10)
at async T0 (factory.ts:140:10)

and I now have a zombie tab which is impossible to close.

I can avoid creating a zombie tab by clicking "Edit" and "Save" manually in the form. Even though there are no changes, that apparently syncs the state properly.

Requesting the object with AML studio, newly created items in potential zombie state look exactly like ones which have been edited and saved, so the problem isn't in the database end. Fully logging out and logging in again also clears potential zombie state, but merely closing the Part does not. So it's some kind of Item cache in the frontend which thinks the Risk object is dirty - or possibly it's the Form Item which is dirty.

I suspect it has something to do with how I save the item. It's saved in a slightly unusual way, in a popup form. But before I dig into explaining how, does any of this sound familiar to you? Perhaps someone with better understanding of how the front end caching in Aras works can immediately tell me what's wrong here.

This is Aras 31 by the way.

Ask A Question I

Thu, 21 Nov 2024 11:02:30 GMT

Ask a question in a forum.