I am doing "traditional" lotus notes programming (same since R5) and need to implement linking between 2 document types (forms) residing in different databases.
Document of type (A) in database (A) can reference several documents of type (B) in database (B).
And document (B) should also display its relationships with document (A), as document (B) can be related to different documents (A).
We have Many to Many relationship.
At the moment it is implemented on one side only (one to many):
Form of Document (A) contains embedded view of special
"link" documents residing in database A. This link documents are created by lotusScript when user selects documents from database (B). When user clicks on an item in this embedded view, it opens document (B).
Client wants to be able to edit this relationship on any side, so that if he edits it in form (A), form (B) is updated.
Form (B) is supposed to have the same kind of embedded view or a list of associated documents of type (A)
What is the best way to implement it?
Client's infrastracture is Lotus Domino 8.5.2 + Lotus Notes 8.5.2, so theoretically, composite applications approach may be an option.
The reason why I ask this question is that as far as I understand there is no good way in Notes to embed a view from another database.
The requirement is that the database should be present on workspace to be displayed in some sort of dodgy list.
It would be great to be able to specify target database for embedded view by server and replicaID, but instead we have a weird list of random workspace databases.
The main problem is that Notes wasn't designed to handle relationships like that between databases (nor anything besides parent child relationships for that matter). So the solution will have to be a creative one.
A couple of (off-the wall, potentially awful) ideas come to mind. One is to store the references in the documents themselves, and update them whenever the document is saved. That could all be done in LotusScript, and would require searching through the other database's documents to update their references.
Upside is that the performance when reading the documents would be excellent. There'd be no issues while reading Database A if Database B was unavailable. It keeps data local to each database. The downsides include the likelihood of save conflicts and the danger that references could get out of sync if documents aren't "saved" but instead are updated via agents, etc.
Another thought is to use agents to manage the links on a scheduled basis. If you don't need real-time up-to-date references, you could run an agent that scans Database B and updates the references in Database A. With this method you could choose either to update the Database A documents themselves - or - as it sounds like you've already done, create a set of link documents that show up in an embedded view. The latter eliminates the save conflict problem.
One more idea is to hide any references when you open a document in Database A, but provide a button to "show" or "update" references. When you click that button, it fires off LotusScript to search Database B and build a list on the fly. This would probably work quickly with less than 10,000 documents. That function could update the link documents you store on the same database which feed the embedded view.
Hope this helps!
What is the best way to implement it?
As you mention creating a composite application may allow you to do this, but would be restricted to windows rather then design level in the form.
eg.
[Window A] --- trigger ---> [Window B]
If you are not familiar with this system I did a tutorial which explains the basics.
http://www-10.lotus.com/ldd/compappwiki.nsf/dx/ibm-my-first-wire
Although the tutorial calls back to the same database, it is easy enough to point to a different one.
Personally I'd do it through XPages. I personally find it much easier to implement then through classic style notes design/comp apps. It will also allow you to display design elements within the same screen area.
As you've already heard, Lotus Notes has no referential integrity constraints built-in, you have to do it yourself.
I wouldn't be relying on document links as they're geared around UNID's which can change if you cut and paste the same document, thus losing the link. Try this,
1/ Create an "ID" field on each document. You can populate it by using #Unique in a computed field to generate an ID, and save that to the documents in both databases. You can create an agent to do this in lotusScript (LS), or formula. (Consider using the evaluate statement if doing in LS)
2/ Create a lookup view in each database that lists the documents by the new ID (don't forget to set the "sort order" of the ID column.
3/ Using an action button that can be configured for both databases, you can create a LS function that will open the the opposing databases view and return the ID field. (NotesUIWorkspace.pickliststrings would be the simplest way to pick the documents, otherwise you could build a dialogbox. Store the list of results in a field called "linkedID" as multi-value list.
4/ There may be more info that you want to store like document title or author, so you'll need to then get a handle to those documents using getdocumentbykey and then interrogating the fields you'll need to display information on screen.
5/ You can then also add a new field on the target documents you're referring to, call it "referrerID", which is a list of documents that reference the current document. This will maintain the two-way relationship.
The field that stores links must be a multi-valued field, otherwise it gets quite cumbersome to loop through list of linked document ID's and manage them.
This approach uses a static key so you can copy databases around without losing the relationships between documents the user has invested time in producing. You can (and probably will) lose those relationships if you rely on document universal ID's (described well in the #documentUniqueID documentation), if you cut and paste the document, or copy the database somewhere else they become new documents despite copying the same fields, and will be assigned a new universal ID, any document links for the old document will be invalid.
If the information you're displaying from the other database changes, you'll need to be able to refresh that data regularly, so consider writing a scheduled agent that can do the look up and refresh the relevant fields.
If the user intends to un-link or change the relationships between documents, then you'll need to add functions that loop through the key fields and keep the lists consistent with what the user is doing. So, like I said, Lotus Notes' flat data structure requires you manage all integrity constraints yourself.
If you want to get a little fancy you can use embedded views as they do support references from another database on the same server. Some tips about handling it in LotusScript here. And use an additional view that categorises your data by the referring ID. Embedded views are ok, as long as the view they're based on is not too big, otherwise it may affect the performance of the form that it is embedded into.
Related
Hoping someone can point me in the correct direction for an XPages application we are writing inside the Domino Client (Notes?) viewer.
I have a view of documents which is being returned, this view has categories on it, and shows fine as this in an XPage, we now apply a filter to the view to limit it to specific owners of the documents, but as soon as we apply the filter, the categories disappear, which means we are left with a long list of documewnts, but unsorted - is there any way to display a filtered view in a categorized manner, on an XPage.
Moving further down my list, I also need to be able to select these documents (and one or many owners) to send to an Lotus Agent which will then create a JSON document to be sent to our friends at DocuSign requesting signatures from the selected owners on the selected documents. I'm not sure what an Agent is yet, but that is the goal ...
Caveat: I'm not a Domino developer, so excuse me if some of the terminology is incorrect.
Categorised views are a very "Notes" construct. When you filter a view, it will only show documents, but not categories. While they are practical in the back, they are cumbersome in the UI.
There are a few design considerations how to tame them in a webUI. However if your users love them, you might consider to flatten them out and recreate the categories in the UI (client side) only.
The actual better way for your use case: add another view that is firstly categorised by the owner and secondly by your category. Use the category filter of the view control to limit the documents to that author. This should do the trick. Eventually use one of the controls from the extension library.
For the agent: don't bother, that's "old Notes speak". An agent would be a piece of code (LotusScript or Java, but since you do web interaction: Java) that gets triggered by an event: manual, on schedule, on document create/update (with some delay).
Since you are in an XPage, you have easier options at your disposal: create a Bean that has the JSON format you need, add a method that takes a Notes document as parameter to populate it, something like public void populate(final Document doc) {...} and use e.g. the GSON library to simply marshall them to JSON (or a collection of them). The GSON library probably is on a current Domino, I put it there as part of VoP 1.0.
Then use a managed bean to talk to Dokusign. When traveling down the managed bean road is is much easier to test than trying to mess with agents.
Hope that helps and ask more questions! (Check the Learning XPages Cheatsheet too)
I have an xpage that allows the user to choose a customer and then order products for that customer. It's not a simple xpage that created a document, uses a view control to view it and re-edits it. It will be used on the web and in the client. How do I fill in all the data for the various fields when the user wants to look at their order for a company since there are multiple documents that make up that xpage? Is there automatic processes or do I need to do it manually?
The best method is to use multiple datasources (the Notes documents) each with a different datasource name. When saved, be sure to save each of the datasources that have a change. Also, it is helpful to mark to "ignoreRequestParameters", so each one acts independently.
I have found that using the dynamic content control useful when doing things like this, it seems to reduce the number of replication/save conflicts.
I am developing a database on Lotus Designer 8.5 environment with LotusScript and LotusFormula.
Is there any possibilities on get a list of recently opened documents (for example last five) in a Lotus Notes database ? My purpose is to provide an embedded view showing the recent documents opened by current user on the current database, which will act like some kind of history view.
Please advise
Use a folder categorized by username. Add the current document in PostOpen event to folder with document.PutInFolder(folderName) and remove the oldest document from folder with document.RemoveFromFolder(folderName)
This way you don't need to edit the documents and can show the last visited documents for a user in an embedded view.
As an alternative you can use user specific folders with option "Shared, private on first use".
I have that functionality in one of my databases.
I just added some code in the QueryOpen event of the form to store the UNID of the document in a profile document linked to the specific user. The values are stored in a multi value field, and my code removes the oldest entry when the number of entries I want to store is exceeded.
The user can actually set that number themselves in teh applications settings, 5 is default but they can make it more or less.
I built a class for this, makes it very easy to modify later, and to implement it in different forms, for different document types.
I then built a method to expose the last documents to the user, using a dropdown box as you can see below. Since you only wwant/need the five (or perhas ten) last documents, no need to use a view.
I need to create some functionality in our SharePoint app that populates a list or lists with some simple hierarchical data. Each parent record will represent a "submission" and each child record will be a "submission item." There's a 1-to-n relationship between submissions and submission items. Is this practical to do in SharePoint? The only types of list relationships I've done so far are lookup columns, but this seems a bit different. Also, once such a list relationship is established, then what's the best way to create views on this kind of data. I'm almost convinced that it'd be easier just to write this stuff to an external database, but I'd like to give SharePoint a shot in order to take advantage of the automated search capabilities.
Proper Parent/Child in Sharepoint is near impossible without developing it yourself. There is one approach to that here: Simulate Parent / Child relationship in SharePoint 2007 with Folders & Content Types
(Note: This concerns SharePoint 2007. In 2010, Joins make this much easier)
Do it in a separate database, create a page(s) with controls that surfaces the data and run search over that. Loses quite a bit of the SharePoint features though.
Otherwise it may be okay to create a custom field control that will allow you to lookup the data in the other list.
The custom field control can be the one to "view" the related data.
I know we have done it for parent child relationships between pages on the same list. Not 1-to-N though.
Tough choice either way.
My vote is "to write this stuff to an external database"
You miss a lot of things in Sharepoint things like transaction support, referential integrity, easy way of updating (compare SQL), reporting (using Reporting Services and a SQL database)... see sharepoint as a way to store documents and simple lists.....
The argument for Sharepoint is if it is a small application, no requirements on support for transactions, no need to import external data etc...
When people say Sharepoint is a development plattform there is a need to define whjat they think a development plattform is.
The latest rumours about Sharepoint 2010 tells us that there will be support for SQL server based lists in next version ..... which I think will at least move Sharepoint in the right direction ....
Take a look at SLAM, SharePoint List Association Manager, an open source project my company created and actively supports. SLAM allows you to synchronize SharePoint data to SQL, including any relationships between lists. SLAM, in addition to being very useful on its own, is really a framework intended to allow developers to create their own complex data associations using what we call SLAM type profiles. We have one out-of-the-box type profile which is part of the open source project which actually allows you to make a SharePoint list hierarchical using the nested set model. For more information, see this page on our codeplex site.
I do this a lot just using sharepoint, using a framework called AAA (Activity,Assignment,Artifact), which allows you to use lookup columns to link an assignment or artifact to a parent Activity. You then build a web part page with connected web parts that allow you to filter all assignments and artifacts by activity. For example, click next to a submission in the submission web part, and all of the submission items attached to that submission will show up. Works great.
The other approach that you can look at using is persisting XML with a field in the item. This is the approach used by the Podcasting Kit (on CodePlex) to store things like ratings.
One possible method is to create a submission content type based on the folder content type and a submission-item based on item content type. Then you can store data hierarchically like in file system and also will work default views and search functionality.
Other way is to create lookup field that points to same list (list=”self”). This field will be used like reference to parent item and you will get list that contains recursively related data. To use this data programmatically will be ok but using views functionality will be little bit complex.
It's easy to do using a connected web part.
Create two lists:
Parent (Id, Title)
Child (Id, Title, ParentId)
Create a new sharepoint page, add DataFormWebPart (displaying Parent) and another one for Child, set both of them to filter based on a QueryString parameter (use that Parameter to filter Parent.Id, and Child.ParentId) voila, you can display parent-child relationships. Now, adding children is more difficult, and that's the part I haven't worked out yet.
Is there a standard way to associate custom properties with a user? I need to store the number of items per page a user selected in a grid of a document library separately for each user and document library.
Edit:
Sorry about this vagueness, I wanted to do it programmatically. It seems like I've found the solution, it is UserProfileManager class, though I'm now looking into whether there is a limitation on the number of properties you can save this way for a user, because the easiest way of saving page sizes on per user+document library basis seems to be using GUIDs of Views as property names and numbers of pages as values. Though I don't know if it is more efficient or not, depends on how sharepoint stores these properties.
No, you would need to create custom code to store the data.
Given the potential amount of data created, it may be wise to store it in a separate database.
This would give greater flexibility in the way the data can be manipulated and retrieved.
Your question is a bit vague. Are you looking to do any custom code? You could do this many ways so it is difficult without knowing more of what you want.
Using custom code you could set up a workflow or event handler to respond to item events and record the information and store it using the User's Profile or as an SPPersistedObject.
If you want a less developer centric way to do it you can use auditing and simply do reporting on your audit results.
You could set up a list to store the selection data, then use events/AJAX on the document list to push tick/untick items into the selection list (store user, library and document as a minimum.
If you don't want a separate list, you could create a field in your document library that stores which users have a given document tagged... You'll still need some kind of event/AJAX to update the list when a user ticks/unticks the box. Crude :)