In Table 3
note that we get a "state_ab" result for every state, as well as the
coordinates needed on our map, but for those states that are only found
in "tbl_states" and not in "biocenters" we get a row with no data in
the "city" field, a condition that we'll use to trigger a different
result than those that do have cities defined.
Our next group
level defined by the ORDER BY of the query, and called by the "group"
attribute, is "status," and, finally, "city," which we'll call twice,
once nested inside of each of the two possible values for "status."
You can see how this information is used by following the program
listing for "MainMap.CFM" abridged from the actual site to include only
the code we need to illustrate the process used to populate the map
display (see Listing 1).
The first item of business once we've called the query is to offset our
map coordinates to account for the spot on the page where the
containing div will display it. Then we begin a rather detailed grouped
output that's used to set styles for the items that will display on the
raw map.
Because we used a left outer join in our "getmapinfo" query, at the
level of the group="state_ab" the parameter #state_ab# will be a list
of the 51 states (including the District of Columbia) whether or not
there are join matches. We'll evaluate each of these rows to see
whether there's a city defined (because we asked for "SELECT distinct,"
those cities that have more than one match will be listed only once).
The rows with no city defined (meaning there are no plasma centers in
that state) are used to create one kind of div, "#state_ab#_set," and
those that do have at least one city defined have created for them some
additional offset information, the div "#state_ab#abbrev" and the
additional "div #state_ab#detail."
That first <cfoutput>, in the <style> section of the
<head> block, used only the topmost group of the query, since all
we needed to do was list the states and evaluate if, based on whether
they had data in their #city# fields, they needed merely the div that
will write the abbreviation in place on the map; or, if they did have
centers present, the richer pair of divs that we'll use to provide
rollover display and links.
Digging in at All Three Levels
In the body of the
page, however, we use a <cfoutput> that employs the "group"
attribute at all three of the levels offered by the query's "ORDER BY"
clause.
Once again, the <cfoutput> begins by being grouped on "state_ab"
and we evaluate whether a given #state_ab# has city data associated
with it, our tip-off that there's at least one plasma center in that
state. For those with no city, we write the div that writes the state
abbreviation in white onto the map and we're done with that state.
But for those states with city data, we write a link that will pass the
#state_ab# parameter to a page that will show the detailed information,
and then after we pull the #state_full# state name to head the rollover
div, we have a nested <cfoutput> that is grouped on "status," so
we can differentiate between centers that are already open and those in
the preparation stages. Here's a case in which we use the next level of
the <cfoutput> grouping twice, once for each of our two possible
values of "status."
The result is a map that highlights those states that have biocenters
in them, which provides those states with a link to a detail page for
drilling down and summarizes the centers in the state grouped by open
status in a rollover div that shows when the mouse hovers on the link.
Grouped Newspaper Columns
I'll pass along one more example of the use of the "group" attribute that I find particularly useful (Figure 1). There are times that I like to display query results in newspaper-like columns (Table 4).
Because this grouping of the data all happens on the output side, we
can operate the logic that determines how a column will be distributed
at whichever levels of the nested <cfoutputs> that helps us make
an attractive display. In the case of listing the states and cities
with biocenters in them, I wanted the cities to stay grouped with their
states. First I set up the query "GetSoonAndOpenCenters" with the ORDER
BY state_full, city clause, a sub-query that counts the distinct
instances of state_full, and output it as shown in Listing 2.
Note that rather than use the query's rowcount as the trigger for
jumping to a new column with the <td> tag, we set up a
incrementer based on the "state_full" count and put the evaluation
trigger for the column change at the end of each iteration of that
outer <cfoutput> group. The result is a reasonably evenly
distributed three-column display that keeps the cities grouped in their
state.
In all of these examples, the utility of the "group" attribute rests
with the fact that it resides on the output side of our data handling,
giving us multiple ways of distributing and displaying the results of a
given query. This includes adding styling, conditional branches in the
various nested output levels, and the association of our choice of
styling with each sublevel of the output. Sometimes it reminds me of my
dear mom, who used to stress to me in my school years that getting in
with the right group was the key to success!.
Subscribe to ColdFusion Developer's Journal Email News Alerts
Subscribe via RSS
