I am building a Booking model in Rails 3.2.3 where a user steps through several form-screens of choices. If the form exactly mirrored the underlying model I know I could use a gem (e.g. Wicked gem) to build a multi-step form. The issue I am having is that on one of the form-screens there are multiple options for the user and within each of those options there are multiple options coming from an external web-service. In other words, on step 2 of the multi-step process, the form presents the user with many options of Rates from our database and for each Rate we have multiple additional options from the web-service. So a user would need to choose one of the options from the web-service (radio buttons) and then make a selection of their Rate of choice. This is then repeated multiple times on this page (although the user can only choose one radio-button option from the web-service and one Rate).
Where I am unsure of best practice is that I can display the multiple options from the web-service as radio buttons but there is a Hash of values associated with each of those options and hence with each of those radio buttons.
So, the question is, should I be attempting to pass that Hash as a param to the next step of the form process or should I be making that into an object and passing that or something else entirely!
I know this is a long explanation but I feel it's a critical point in the design of this workflow and I want to get it right.
Many thanks in advance,
J.
EDIT
Thinking it through again, the initial problem is how to represent a series of radio buttons when each radio button represents many values as opposed to say an id (in this instance each radio button represents a hash of values from the external web-service). Should the hash be made into an object and this passed instead - something along those lines?
I figured this out. On inspecting the "hash" coming back from the external web-service, I noticed that date fields were not in quotes, e.g.
:departs=>Wed, 21 Aug 2013 10:40:00 +0000
whereas all the other fields were in quotes. This made me a little suspicious. So in the end I used to_json on the returned hash:
response_hash_from_webservice = data.to_json
In the form I then used:
JSON.parse(response_hash_from_webservice).each do |nested_item|
# Access elements like so
... nested_item['company_name'] etc ...
end
However I needed to post this nested_item through to a next step of the form (as a radio button) and it only worked by again using to_json.
<%= radio_button_tag 'nested_item', nested_item.to_json %>
I could then post this value or put it in the session and on the following form page use:
require 'json'
hash = JSON.parse session[:nested_item]
And then access the values as normal:
<%= nested_item['company_name'] %>
Related
I am a research student working on an access database where I have created a continuous form that is to be used in a sort of scrolling electronic health record format. The header includes information about the patient and the continuous form aspect is a series of patient visits. In the header, I have a few subforms, which populate based on linking on the patient_ID number which is posted in the header (generated by vba macro such that previously reviewed patients aren't revisited). These subforms seem to significantly lag such that they the results from the previous patient stick around going into the next one. The subforms contain user-selected pertinent data. Each form has its own table. They are linked primarily based on patient_ID.
I have tried:
macro on the header detail: on click, refresh. - seems to work, but not very elegant/intuitive
macro on the main form - same as above but on load, click, got focus, lost focus, open, activate -- none of them seem to do anything.
forced requery via vba (see below) on opening/etc of the form. Neither way has worked. Tried to run these on opening the header form.
Public Function RequeryMain()
Dim frmMain As Form
Set frm = Forms("FRM_continuous_reports_patient")
frm.Requery
End Function
Public Function RequeryHeader()
Dim frmHeader As SubForm
Set frmHeader = Forms("FRM_continuous_reports_patient").FRM_continuous_header_working
frmHeader.Requery
End Function
In the end, it is frustrating for users to have to click to clear the form for new entries. It works otherwise.
The end goal is for the form to open and have all the subforms load based on the newest patient_ID. This would likely have to involve a staggered load: (1) VBA script selects the next patient based on certain characteristics and passes the patient_ID to the main form; (2) load main continuous form based on patient_ID submitted to it; (3) load the header subforms and any pertinent data within (although should be blank for the first time these are seen); (4) on completion, back to (1).
From what I understand, this is already how it is working, however the subforms are loading too quickly? How can I fix this?
Hopefully someone can help explain how to remedy this/correct any misunderstandings I have about the mechanics of forms.
I know this will sound odd, but subforms actually load before parent form. Lag in subform display is not something I have encountered. Code should not be needed and likely will not correct. Must be something about the form/subform design, maybe their RecordSource. Would have to examine db to determine.
It is not necessary to create form objects in VBA just to requery. Is code behind the main form? Me.Requery will be enough for the main form. I always give subform container a name different from the object it holds, like ctrPatient. Then just Me.ctrPatient.Requery.
Why not put subforms in Detail section?
To answer your questions, and provide a condensed version of the logic referenced at: https://accessexperts.com/blog/2014/01/07/delay-loading-subforms-in-access/
is to:
In design view, set your subform SourceObject to "" (and save your form)
When you are ready to show the subform, just execute: Me..SourceObject = ""
When you are ready to navigate to the NEXT patient, clear the subform link: Me..SourceObject = ""
Now that should solve the issue of out of sync data between the main and sub.
You don't need to use the CASE statements, but they operate as if you had a bunch of "If" / "ElseIf" all nestled together -- except the CASE makes it easier to follow. It basically gets a value from a variable (i.e. Select Case MyVariable); then checks to see if it equals what you want (Case 1, Case 2, etc.), and if so, does whatever you code.
I'm trying to use Adobe DTM to pass the date to a prop variable but haven't had much success. The final output should be a prop report in Adobe that'll provide me traffic data for specific dates (5/11/16, 6/15/15 etc). The ultimate goal for setting the dates as a prop is to be able to classify a range of dates based on various business needs.
Could anyone point me in the right direction for getting this done? I am assuming I'll have to add a line of code in the s.code file that'll define s.prop5 = ...
Thanks
Based on your comments, it sounds like you are just looking to pop something with the current date stamp with "MM-DD-YYYY" format.
As Gigazelle mentioned, you can create a Data Element to return the value, and then reference it for setting your prop. However, throwing a data layer into the mix may be overkill for you, depending on your needs/limitations.
Data layers are meant for exposing data that DTM can't feasibly/reliably automate on its own via built-in features or hosting an autonomous js snippet.
The only reason you might want to consider having your site push it to a data layer, is if you want to generate the date via server-side coding to ensure the date is generated within the same timezone setting for all visitors. If you generate it client-side, it will be generated according to the visitor's browser/system settings. Since visitors are from all over the world in different timezones, the data may not be as consistent (even if you add additional code to change the timezone offset, it still may not be 100% based on browser version/security settings, or visitors who alter their browser/system date/timezone settings).
So, if you want to ensure the best accuracy, then I suggest you output the value via server-side code, and put into a data layer. How you do that depends on your server and what language you have at your disposal for your web pages being served up. Here is a very basic example using PHP:
<script>
var dataLayer = {
currentDate : '<?php echo date('m-d-Y'); ?>'
}
</script>
This will have the server generate the date stamp and output a js object called dataLayer with a property currentDate you can reference. You can create a Data Element as Type "JS Object" and for Path, put dataLayer.currentDate, and then reference your Data Element elsewhere (see below).
If that's too much trouble for you or if you want to keep it pure client-side/DTM and are okay with the potentially lower consistency...
Within DTM, go to Rules > Data Elements, and click Create New Data Element.
Name it "currentDate" (no quotes).
For Type, choose "Custom Script", and click Open Editor, and add the following:
var t=new Date(),d,m,y;
d=t.getDate();
d=d<10?'0'+d:d;
m=t.getMonth()+1;
m=m<10?'0'+m:m;
y=t.getFullYear();
return m+'-'+d+'-'+y;
Click Save and Close and Save Data Element.
Now you can reference the data element to pop prop5. How you do it depends on how you've setup Adobe Analytics within DTM. For example, if you set it up as a tool and only want it to pop on initial page view, you can open your AA tool config, go to the Global Variables dropdown, and set prop5 there. You reference it as %currentDate%
You can do the same %currentDate% syntax in a Page Load Rule or other rule or any other place that uses DTM's built-in fields.
Alternatively, if you need to reference it within javascript code (e.g. if you are setting prop5 within s.doPlugins or some other Custom Script box, you can reference the data element like this:
s.prop5 = _satellite.getVar('currentDate');
Set a JS variable on your site (such as within a data layer) that outputs the date in the format you wish to collect it in. Something like var d = Date();
In DTM, go to Rules > Data Elements and create a data element that maps to the JS variable you created on your site
If you want prop5 to be defined on every page, click the gear icon and map prop5 to your data element name by using %DataElementName% (whatever you named your data element in step 2, wrapped in percent signs). If you don't want it defined on every page, go to Rules and create a page load rule, event based rule or direct call rule depending on when you want the variable to trigger. Under the Adobe Analytics section of the rule, map prop5 to %DataElementName% .
This will allow you to collect dates as values, which can then be used in classifications.
I have a window with a search form at the top and grid at the bottom.
User can enter values in the search form and click button - Get Records.
At the click of this button, I load the store of the grid by passing the values in form fields as parameters in following way:
store.load({
params:{
key1:Ext.getCmp('field1').getValue();
}
});
I tried giving parameters in the store proxy itself, but it unfortunately always takes up initial values (values when the form is rendered) and not the latest one entered by the users in the form fields. Following is the method I used for assigning values to params while creating the store:
extraParams:{
key1:Ext.getCmp('field1').getValue();
}
I wanted to seek guidance at two things:
a. While defining a store, can I ensure that store takes latest/current values from the form fields before querying server, so that I don't have to provide these values while calling load function?
This becomes more necessary as I have a paging toolbar at the bottom which carries a refresh button (along with next, last, previous, first icons for navigation).
Now, whenever user clicks at refresh (or any navigation icon), the store gets loaded without the query parameters.
Thus the second thing is:
b. If the answer of 'a' is that - Pass the latest values to parameters manually when calling load function - then how can I write the handler for 'refresh' button and navigation icons (that is, next, last, previous and first) in the paging toolbar, so that I can pass the latest form values to load function.
Thanks for any help in advance.
PS: I am using ExtJS 4.
yourStore.on('beforeload',function(store, operation,eOpts){
operation.params={
status:cmbStatus.getValue(),
value:txtBuscarPor.getValue(),
empresa:'saasd',
app:'dsads'
};
},this);
Related to your question (b) (and because you especially asked for this in the comments section):
I see only one standard way to hook into the PagingToolbar button handlers which is very limited.
Ext.toolbar.Paging fires a 'beforechange' event before it actually changes the current page. See API docs. A listener that returns false will stop the page change.
All other methods require extending Ext classes which wouldn't be a problem if the ComboBox would make it easier to use your own implementation of BoundList (which is the class that renders the dropdown) or pass through config parameters to BoundList resp. the paging toolbar.
I tried to bring this lack of flexibility up on the Ext message board once but was pretty much ignored.
A possible solution for this is to use 'beforeload' event of the store and provide the list of parameters in it. This way, whenever the store is loaded, then its beforeload event is fired and the values picked up are always the latest. Hope this helps someone looking for something similar.
What is the standard method for implementing a "wizard" using successive web forms?
I'm implementing a CGI that accepts several options, files, etc. But some of these options have dependencies to one another, and allow or require other options to be used.
For example, one type of object that needs to be initialized by the CGI can be created using:
two files of type X
two strings
one file of type Y
In my command line version, I look whether two files of type X, two strings, or one file of type Y is provided, and construct the object in the appropriate manner.
In my CGI, I'd like to do this using multiple pages or DHTML (perhaps a radio button that specifies which arguments the user wishes to provide; changing the radio button will change the form to the right).
Anyway, I have this situation for 3 main groups of arguments. I thought it would be pleasing to the user to create a 6 "page" wizard (think online dating):
Page 1:
"How would you like to specify your proteins of interest?"
radio button:
Two FASTA files
Prefix and suffix strings that match all of my proteins (and match only my proteins)
A text file containing the proteins
Page 2:
"Great! Please choose your (either 'fasta files', 'prefix and suffix strings', or 'text file')."
(appropriate web form)
Unfortunately, if the form is split over different pages, I'm not sure how the 3rd, 4th, etc. pages will know the location of the temporary folder created for the uploaded files from pages 1 and 2.
I'd really appreciate your advice; I have a good command line app, but I am having a difficult time making beautiful interface code that will do what I want. And I'd be shocked if there isn't a very easy standard way to do this with Django or some other framework; it just seems it must come up very frequently.
There's a wizard plugin for jQuery.
http://plugins.jquery.com/project/formwizard
If you don't know jQuery, it is a javascript framework for doing DHTML.
Try the demo at http://thecodemine.org/
I have a three step form where each step posts to its own action. The action redirects to the next step. The data is stored in the session scope. I have a filter that prevents a user from accessing the form handlers through anything other than a post request.
There's nothing to stop someone from manually typing in the address of a step, however. To deal with this problem I set a currentStep variable in the session.
<!--- Some data is processed here --->
<cfset session.currentStep = "stepTwo">
And in step two I would check for a structkey:
<cfif NOT session.currentStep = "stepTwo">
<!--- redirect to #session.currentStep# --->
This approach works, but it has a major drawback: A user can not press the back button in the browser window, or edit any data he or she has already entered.
What are some the best practices to implementing a multistep form? Can I improve my process to incorporate back-button functionality?
Instead of using the session variable to only allow them to access the current step, allow them to access the current or previous steps. Sort of a "how far you can go" flag.
Now, add links to the previous steps, like a breadcrumb trail.
Finally, use a lookup in the persistent store (db, session, xml, bag of holding, etc.) for the data already entered for that form. Create a blank set of form data, overwrite it with anything found in the persistent store, then overwrite it with anything from the form scope itself. Something like:
populate = structNew(); // this is the data to populate your form with on load
populate.someValue = "";
structappend(populate, dataFromStorage);
structappend(populate, form); // from things submitted from the form scope, in case validation fails
<input type="text" name="someValue" value="#variables.populate.someValue">
Now, if someone hits the same form step twice, they will see (in order of precedence) the values they submitted, but which didn't pass validation, values from the persistent data store, and then an empty form.
You can stay using Session approach if you want.
To solve your major drawback, you can change your logic a bit.
At the last step, make sure data of all steps are found in the session. If not, redirect the user to the first unfilled step? Shouldn't be too hard.