转：http://www.infopathdev.com/blogs/greg/archive/2005/07/14/Create-a-Repeating-Cascading-List-from-a-Flat-Data-Source.aspx
Create a Repeating Cascading List from a Flat Data Source
It
is common to need two or more list boxes where the choices in one are
dependant on the value of another. This is known as a cascading list. A
cascade can occur with as few as two levels up to any number of levels.
Some common scenarios which require cascading lists are:
- Location: State / City
- Vehicle: Make / Model / Trim
- School: Year / Semester / Class / Instructor
As
the number of levels in the cascade increases, so does the complexity
of the process and work involved. Also, if the cascading list is
presented inside of a repeating table or repeating section you must
account for another issue: data crossover. Data crossover means that
the options presented in a cascade on a given row will be affected by
earlier selected values from other rows—thus destroying the true
cascading effect.
In this task we will create a flat secondary
data source containing the values used to populate our lists. We will
then create a three-level repeating cascading list using drop-down list
boxes. We create three levels of cascade to reveal the compounding
complexity involved for level three and beyond. Because of the nature
of the flat data source, we will filter each list to ensure that only
unique values are displayed. We will resolve the issue regarding data
crossover using the current function. And finally, we ensure
valid selection combinations by clearing selected list values as
selections earlier in the cascade are changed.
THE FLAT SECONDARY DATA SOURCE
Flat
data sources are commonly returned from databases, Web services, and
SharePoint. A flat data source has no hierarchy in the structure; it is
simply a list of identical elements, each with an identical set of
attributes. Only the values of the attributes change. Flat data sources
also require additional filtering to ensure that only unique values
appear in the drop-down lists. Let's start by creating the flat
secondary data source.
Create the flat secondary data source:
Copy the following code into a text editor, and then save the file as Vehicles - Flat Data Source.xml . This file will be used as a secondary data source in our form.
|
<?xml version="1.0" encoding="UTF-8"?> |
THE FORM
Now
that we have our flat secondary data source we are ready to create our
form. First we will design a simple schema to support a three-level
repeating cascading list. Then we will design the layout using
drop-down list boxes in a repeating table. When the layout is complete
we will focus on creating the repeating cascade.
Add the secondary data source to a form:
- Design a new blank form.
- Choose Data Connections from the Tools menu, and then click Add.
- In the Data Connection Wizard, select Receive Data, and then click Next.
- Select XML Document, and then click Next.
- Click Browse, locate and select the Vehicles - Flat Data Source.xml file, click Open, and then click Next.
- Click Finish, click Yes, and then click Close.
Create the schema:
- Open the Data Source task pane.
- Add a Group named Vehicles .
- Select the Vehicles group, and then add a Repeating Group named Vehicle .
- Select the Vehicle group, and then add three Text Fields named Year , Make , and Model .
Design the view:
- Right-click Vehicle , and then choose Repeating Table.
- Select all three header row cells, and then set the font size to 8 .
- Right-click each of the three text boxes, and then choose Change To | Drop-Down List Box. The results are shown in Figure 1.
Figure 1. Layout of the repeating cascading drop-down list boxes.
- Double-click the repeating table label.
- On the Data tab of the Repeating Table Properties dialog box, change the insert button hint text to Insert Vehicle , and then click OK.
THE CASCADING LISTS
With
the form layout complete we will now turn our attention to the meat of
the task: making one list cascade off of another. The process is not
very difficult, but can become increasingly complex as the number of
cascade levels increases. Creating a cascade requires the use of
filters. The options displayed in one list box are filtered based on
values selected in earlier list boxes. Because our cascade is in a
repeating structure, we must use the current function to
ensure the filter does not crossover to values in other rows. Let's
populate each drop-down list box using our secondary data source and
create the cascade filters.
Populate the Year drop-down list:
- Double-click the Year drop-down list box.
- In the Drop-Down List Box Properties dialog box select Look Up Values In A Data Connection.
- Select Vehicles - Flat Data Source from the Data Connection drop-down list.
- Click the Select XPath button to the right of the Entries text box, select Vehicles/Vehicle/year , and then click OK twice to close the open dialog boxes.
Populate the Make drop-down list:
- Double-click the Make drop-down list box.
- In the Drop-Down List Box Properties dialog box select Look Up Values In A Data Connection.
- Select Vehicles - Flat Data Source from the Data Connection drop-down list.
- Click the Select XPath button to the right of the Entries text box, and then select Vehicles/Vehicle/make .
Filter the Make drop-down list based on the selected Year in the current row:
- Click Filter Data, and then click Add.
- In the Specify Filter Conditions dialog box, select Select A Field Or Group from the first drop-down list.
- In the Select A Field Or Group dialog box, select Vehicles/Vehicle/year , and then click OK.
- In the second drop-down list, select Is Equal To.
- In the third drop-down list box, select Use A Formula.
- Type the following formula (as shown in Figure 2):
|
current()/my:Year |
Figure 2. Filtering to compare the Year to the value on the current row.
- Click OK five times to close all open dialog boxes.
We
mentioned you would encounter the issue of data crossover. An example
of data crossover is shown in the difference between Figure 3 and 4
(please note that Figures 3 and 4 already account for unique value
filtering). Without using the current function the vehicle
make will be filtered on all years currently selected in the repeating
table, presenting the user with invalid choice combinations. Thus if
you have 3 rows with the years 1963, 1979 and 2004 selected, each of
the Make drop-down lists will contain a list of makes for all
three of those selected years, as shown in Figure 4. By properly using
the current function, we avoid data crossover by limiting the choices in the Make drop-down list to only those of the selected Year on the same row, as shown in Figure 3.
Figure 3. Vehicle makes filtered by year using the current function.
Figure 4. Vehicle makes filtered by year without using the current function.
At this point you have created a two-level repeating cascade. The options in the Make drop-down list are filtered based on the selected Year
in the current row. You can test this by previewing your form. You will
notice that each drop-down list contains repeated values. This is a
byproduct of using a flat data source and will be corrected in the next
section when we filter for unique values. Now let's create the
level-three cascade, which will reveal the compounding complexity
involved in creating each additional level of cascade from here.
Populate the Model drop-down list:
- Double-click the Model drop-down list box.
- In the Drop-Down List Box Properties dialog box select Look Up Values In A Data Connection.
- Select Vehicles - Flat Data Source from the Data Connection drop-down list.
- Click the Select XPath button to the right of the Entries text box, and then select Vehicles/Vehicle/model .
Filter the Model drop-down list based on the selected Year in the current row:
- Click Filter Data, and then click Add.
- In the Specify Filter Conditions dialog box, select Select A Field Or Group from the first drop-down list.
- In the Select A Field Or Group dialog box, select Vehicles/Vehicle/year , and then click OK.
- In the second drop-down list, select Is Equal To.
- In the third drop-down list box, select Use A Formula.
- In the Insert A Formula dialog box, type the following formula:
|
current()/my:Year |
- Click OK.
Filter the Model drop-down list based on the selected Make in the current row:
- Click And to create a second filter condition.
- In the first drop-down list of the second filter condition, select Select A Field Or Group.
- In the Select A Field Or Group dialog box, select Vehicles/Vehicle/make , and then click OK.
- In the second drop-down list, select Is Equal To.
- In the third drop-down list box, select Use A Formula.
- In the Insert A Formula dialog box, type the following formula:
|
current()/my:Make |
- Click OK. The two filter conditions are shown in Figure 5.
Figure 5. Filtering the third level of a repeating cascading list.
- Click OK four times to close all open dialog boxes.
You
have now created a three-level repeating cascade. Preview the form and
test the functionality of the drop-down list boxes. You may discover
two issues: first, the list options are not unique, and second, you can
create invalid selection combinations. We will resolve these two issues
in the next two sections.
FILTERING LISTS FOR UNIQUE VALUES
Values
in flat data sources are usually repeated. Although this is important
given the XML structure, it can lead to option bloat in your list box.
It can only serve to confuse and frustrate the user to present them
with numerous identical options. The answer is to filter the list to
display only unique values. This can be done using XPath.
It
should be mentioned that just as the complexity of the cascade filters
compounded with each additional cascade level, so do the complexity of
the uniqueness filters. You will see this in the expression used to
filter the Model drop-down list for unique values.
Add a uniqueness filter to the Year drop-down list:
- Double-click the Year drop-down list box.
- In the Drop-Down List Box Properties dialog box, click the Select XPath button to the right of the Entries text box.
- In the Select A Field Or Group dialog box, click Filter Data, and then click Add.
- In
the Specify Filter Conditions dialog box, select The Expression from
the first drop-down list, and then type the following expression (as
shown in Figure 6):
|
not(. = ../preceding-sibling::Vehicle/@year) |
Figure 6. Filtering list options to display only unique values.
- Click OK four times to close all open dialog boxes.
Add a uniqueness filter to the Make drop-down list:
At
each cascade level beyond the initial list box, your uniqueness filter
must take into consideration the selection from each of the previous
levels. Due to the nature of this filter, if you do not take the
previous selections into consideration, you may inadvertently
completely filter out all valid options.
For example, in our
sample secondary data source, vehicle years 1987 and 1999 both include
the Honda make. If we did not account for the selected year, the Honda
make would not appear as an option when year 1999 was selected. This is
because the filter would identify that the Honda make appeared earlier
in the secondary data source. The Dodge make in years 2003 and 2004
would also be completely filtered out because Dodge appears earlier, in
year 1987. There are a few other makes and models in the sample
secondary data source that would encounter this issue.
- Double-click the Make drop-down list box.
- In the Drop-Down List Box Properties dialog box, click the Select XPath button to the right of the Entries text box.
- In the Select A Field Or Group dialog box, click Filter Data, and then click Add.
- In
the Specify Filter Conditions dialog box, select The Expression from
the first drop-down list, and then type the following expression:
|
not(. = ../preceding-sibling::Vehicle[@year = current()/my:Year]/@make) |
Note that compared with the uniqueness filter we placed on the Year drop-down list box, there is an additional subfilter specified on the Vehicle node. The subfilter [@year = current()/my:Year] constrains the main filter to verify that we display unique values for each make within the selected year.
- Click OK four times to close all open dialog boxes.
Add a uniqueness filter to the Model drop-down list:
- Double-click the Model drop-down list box.
- In the Drop-Down List Box Properties dialog box, click the Select XPath button to the right of the Entries text box.
- In the Select A Field Or Group dialog box, click Filter Data, and then click Add.
- In
the Specify Filter Conditions dialog box, select The Expression from
the first drop-down list, and then type the following expression:
|
not(. = ../preceding-sibling::Vehicle[@year = current()/my:Year and @make = current()/my:Make]/@model) |
Just as we did with the uniqueness filter for the Make
drop-down list above, here we have added an additional subfilter to
guarantee that we are only eliminating repeated options from the list
of models within the selected year and selected make.
- Click OK four times to close all open dialog boxes.
If
you can guarantee that the values, at any level of cascade, will always
and forever be unique then you do not need to add the uniqueness
filter. If, however, there is any doubt, it is always safer to add it.
For more information on filtering lists for unique values, refer to Display Only Unique Options in a List.
CLEARING SELECTIONS AFTER CHANGES
One
issue with cascading lists that needs to be addressed is that when a
selection is changed in an earlier level the selections following that
change are likely to be invalid. For example, if you select year 1963
and model Chevrolet, and then select year 1973, Chevrolet remains the
selected model, even though our data source does not list it as a model
for 1973. The old value needs to be cleared so that invalid selection
combinations cannot exist. This can be easily corrected with Rules.
With
some extra work, you can probably determine whether the following
selected values are still valid and clear only invalid ones, but there
are advantages to just clearing each following selected value. One
advantage is that rules can cascade as well—like a Domino effect. As
one value is changed, the next value is cleared; clearing that value
triggers the next value to clear, and so on. Rules automatically affect
the current row, so there is no need to filter using the current function in the rule.
The
rules we create will have no condition so that they will be processed
every time the selected value changes. To clear the selected value of a
field, we leave the Value text box of the rule blank.
Clear the selected Make when the selected Year changes:
- Double-click the Year drop-down list box.
- In the Drop-Down List Box Properties dialog box, click Rules, and then click Add.
- Name the rule Clear Make , and then click Add Action.
- In the Action dialog box, select Set A Field's Value from the Action drop-down list.
- Click the Select XPath button to the right of the Field text box, and then select myFields/Vehicles/Vehicle/Make in the Main data source.
- Click OK five times to close all open dialog boxes.
Clear the selected Model when the selected Make changes:
- Double-click the Make drop-down list box.
- In the Drop-Down List Box Properties dialog box, click Rules, and then click Add.
- Name the rule Clear Model , and then click Add Action.
- In the Action dialog box, select Set A Field's Value from the Action drop-down list.
- Click the Select XPath button to the right of the Field text box, and then select myFields/Vehicles/Vehicle/Model in the Main data source.
- Click OK five times to close all open dialog boxes.
Try it:
You
have now created a three-level repeating cascading list that filters
for unique values and prevents invalid selection combinations. Preview
the form and try various selection combinations.