I'm learning about list functions in couchdb, and I'm wondering what the identity function looks like, if one exists. (i.e. a list that emits the same output as a raw view)
If you understand list functions, could you help me write one?
I expect it would look something like
function(head, req) {
while (var row = getRow()) {
send(row)
}
}
but I'm not sure, and that doesn't sound right because I don't use head and req anywhere.
I have an example list function. It's purpose is to do an intersection of tags, you pass in a key (as the view key), and extra_keys which are the additional tags to include in the intersection. I am just telling this as the purpose of this view does not relate to your question, but it will help you understand what you are looking at.
The list mimics view output, mostly. My app only really looks at rows, so that is the only property I add. YMMV. You can cut the cruft in the middle, if you like. The example is here:
https://github.com/ryanramage/eckoit/blob/master/app.js#L209
I found this blog post which seems to be close to what you want. The only difference would be that in the real view results there is a total_rows and offset field (as long as there isn't a reduce function).
Related
I wrote and support a little web app for our local animal shelter to help volunteers locate dogs. I chose Tabulator because it had great features and was easy to use and have been very happy with my choice. For the first version of the app I used external input fields to search and manually did all the wiring to support live search.
Now I am working on v2 and am trying to use header filters. My problem is that the filters need to be exclusive, that is, using filter1 clears/disables filters 2 and 3, using filter2 clears/disables 1 and 3, and so on. With the external search fields I used focus() events to do this. When I try using jQuery on(focus) delegates to do the same with header filters and for example table.setHeaderFilterValue("field1", "") it does not work; the event triggers but the input box never gets the focus so I cannot type in it. I've tried different events like click; but nothing I've tried works properly.
I've studied the docs and struggled with this for several hours. I've considered hooking dataFiltering() and eliminating the filters I don't want, but I'm not sure how to identify the filter that I want to keep, and there is still the matter of the text in the fields to be dealt with. I'm sure it doesn't help that front-end work is not my area of expertise, though so far I've managed well enough. Is there a simple or normal way to do this that I'm just not seeing?
For the record, I found a way to do exclusive filtering with header filters with a single event callback:
\$(document).on("focus", ".tabulator-col input[type=search]", function() {
var hfNames = ["name", "anum", "kennel"];
var fieldName = \$(this).closest(".tabulator-col")[0].getAttribute("tabulator-field");
hfNames.map(function(hfN) { if (hfN != fieldName) table.setHeaderFilterValue(hfN, "") });
});
The three hfNames are the field names of the columns with filtering on. Yes, I could have derived them dynamically, but it didn't seem worth it for a small app like this.
As I suspected, the key was simply a better knowledge of JQuery.
I have two similar problems that I suspect have a common solution.
1) I'd like to create custom Parts that are Attachable, but only to specific content types, only Taxonomies for example. It would be really cool if that was possible out of the box through migrations e.g something like .Attachable(cfg => cfg.ToType("Taxonomy")) but I don't think it is.
Currently, to prevent my custom Part from being used on content that it's not intended for, I just write checks in the driver methods:
protected override DriverResult Editor(CustomPart part, dynamic shapeHelper)
{
if (part.ContentItem.ContentType != "Taxonomy") return null;
return ContentShape("Parts_Custom_Edit", ...
}
Is this a good way to go about it? Would the Handler be better fit for this kind of logic?
2) Similarly, I'd like to be able to conditionally attach different Parts to different individual Content Items. For example, I would like only first level parent Terms in a Taxonomy to have some fields while child Terms have some others.
The best way I can currently come up with to handle this is to just create one Part that holds all fields and run similar checks to the one above in its Driver methods to return different models depending on its container. Then in the template View I check which fields to render:
#if (Model.ThisField != null) {
<div>#Html.EditorFor(m => m.ThisField)</div>
}
else {
<div>#Html.EditorFor(m => m.ThatField)</div>
}
Ideally I'd like to create one attachable Part that's capable of adding several non-attachable secondary Parts to existing Content Items when it is attached to a Type and to new Content Items when they are created or updated. Is there a painless way to do this? I think 'Welding' might be what I need but I haven't been able to find any documentation or tutorials that can explain Welding to me like I'm five.
I think you need to implement a dynamic welding approach. I had to solve a similar issue, it is posted here. Hope this helps.
I have a view in couchDb that is defined like this
function (doc) {
if (doc.url) {
var a = new Date(doc.postedOn);
emit([a.toLocaleDateString(), doc.count, doc.userId], {
_id: doc.userId,
postTitle: doc.postTitle,
postSummary: doc.postSummary,
url: doc.url,
count: doc.count
});
}
};
This gives me the result in a format that I want.Sorted first by date then by count and then by userID.
However I have trouble querying it.What I want is to query this view just by userId.That is leave the date and the count parameter null.
_view/viewName?limit=20&descending=true&endkey=["","","userId"]
does not give me the desired result.
Should I be using list function to filter out the results of the view.Is there any impact on performance if I do this?
This quote from the definitive guide first gave me the idea that list functions could be used to filter and aggregate results.
The powerful iterator API allows for flexibility to filter and aggregate rows on the fly, as well as output raw transformations for an easy way to make Atom feeds, HTML lists, CSV files, config files, or even just modified JSON.
List function has nothing to do with your case. From the docs you've linked to yourself:
While Show functions are used to customize document presentation, List functions are used for same purpose, but against View functions results.
Show functions are used to represent documents in various formats, commonly as HTML page with nicer formatting. They can also be used to run server-side functions without requiring a pre-existing document.
To solve your problem just change the order of the emitted keys, putting userId first, i.e.:
[ doc.userId, a.toLocaleDateString(), doc.count ]
and update your query appropriately.
If changing the order of emitted keys is not an option, just create another view.
Can you "pass" a "parameter" to a view's map function from the query part of a URL? If so, could you show me an example? By, parameter I mean anyway you can access parts of the query string within the map or reduce function.
Thanks in advance.
That's not quite how views works. The view nominates key-value pairs, where the value could be the whole document. You then ask for something where the key is exactly some value or between a range of values. You have to bend your mind around this way of thinking or it won't make sense. The view is independent of parameters, and then it is narrowed down based on the key or startKey/endKey parameters.
I am doing a workflow for a document library. I put a OnWorkflowItemChanged, and I want to get the value of the column which is changed. I use the workflowProperties.Item["name"] and use the afterProperties. But when I use the workflowProperties.Item["column name"], I still got the original value. When I use the afterProperties, it's NULL.
Then I make another workflow that is the same as above for a list. I can use the workflowProperties.Item["column name"] to get the new value in OnWorkflowItemChanged.
Has anyone come across this problem before? Can you give me some help?
The question seems to mix up Item with ExtendedProperties. As to why a difference is seen on a List/Document Lib, it might have something to do with versionining or perhaps the internal serialization is different. Anyway, some of my experience is outline below. I hope it may be of use:
Use the GUID (as a Guid object, not a string) to access the Before / After ExtendedProperties field. Using the Display Name in the ExtendedProperties will not work. The documentation on it is wrong. You can use SPList.Fields to go from Display Name to Column ID (Guid).
I bind all "Before" to MyWhatever_PreviousProperties and all "After" to MyWhatever_Properties, only accessing MyWhatever_[Previous]Properties after the appropriate event(s)).