XML / XSL

8-Apr-04 4:59

i am trying to display some data to the user as HTML. i have been told to convert the data into XML and to then use an XSL style sheet to convert the XML into HTML.

i have the situation:

<specimens>
  <specimen_item>
    <number>1</number>
  </specimen_item>
  <specimen_item>
    <number>2</number>
    <origin>left arm</origin>
  </specimen_item>
  <specimen_item>
    <number>3</number>
  </specimen_item>
</specimens>

i have worked out enough XSL to produce a style sheet that converts my XML into HTML.

the problem is that i want to turn this data into a table. the number tag goes into the Number column, and the origin value goes into the origin column.

the full table has 9 columns, but most of the time there is no data in most of the columns.

in the original C code i am manually generating the HTML from the original data. so i check to see if there is any data in the origin column, and if there is no data i do not build this column.

so i want an <xsl:if> block that only builds the cells for the origin column if string-length(.) > 0 for at least one of the origin tags in the XML.

i am using IE6 to evaluate the XSL stylesheet, and after 2 days i still cannot do this.

setting the current scope to the specimens block, i have tried various variations on:

<xsl:if test='string-length(specimen_item/origin) &gt; 0'>

but this only seems to check the content of the first specimen_item block.

so far the only way i can see to get around this is to add extra tags to the XML that are used to control how the XSL builds the display. if i am going to have to go to this level, then i would rather simply generate the HTML by hand in my C++ code.

any helpful suggestions would be gratefully received.

Philip Fitzsimons14-Apr-04 3:57

14-Apr-04 3:57

you have to do two things...

1) you have to interate of each specimen_item, either by using a for each, or using apply-template. (I would use apply-templates)
2) your test would be better using an element test...

so something like:

<br />
<br />
   <xsl:template match="/"><br />
      <xsl:apply-templates select="specimens" /><br />
   </xsl:template><br />
         <br />
   <xsl:template match="specimens"><br />
      <table><br />
         <tr><br />
            <td>Number</td><br />
            <td>Origin</td><br />
         </tr><br />
         <xsl:apply-templates select="specimen_item"/><br />
      </table><br />
   </xsl:template><br />
   <br />
   <xsl:template match="specimen_item"><br />
      <tr><br />
         <td><br />
            <xsl:value-of select="number"/><br />
         </td><br />
         <xsl:if test="origin"><br />
            <td><br />
               <xsl:value-of select="origin"/><br />
            </td><br />
         </xsl:if><br />
      </tr><br />
   </xsl:template><br />

"When the only tool you have is a hammer, a sore thumb you will have."

feline_dracoform14-Apr-04 4:52

14-Apr-04 4:52

i actually did this with
<xsl:for-each select='specimen_item'>

i havent really got my head around templates yet. unfortunately the templates dont help with the real problem. i only want to create the Origin column if there is data to go into it.

so far i cannot see any way to do this in XSL without introducing a new "is there any Origin data at all" tag into the XML Confused | :confused:

Philip Fitzsimons14-Apr-04 4:56

14-Apr-04 4:56

> so far i cannot see any way to do this in XSL without introducing a new "is there any Origin data at all" tag into the XML

thats what the test I wrote does.... the test will work in either a template or a for-each.

"When the only tool you have is a hammer, a sore thumb you will have."

feline_dracoform14-Apr-04 6:00

14-Apr-04 6:00

*puzzled expression*

i have only been doing XSL for a few days, so perhaps i am overlooking something. still, i think we are talking at cross purposes.

from reading this, and based on a couple of experiments this XSL always writes out the Origin table head item even if there is no data for this column. i am trying to remove the Origin column from the output HTML when there is no data, and i am not seeing this example do this.

Philip Fitzsimons14-Apr-04 6:03

14-Apr-04 6:03

ok understood.
just use the //origin (the // means are there any children called origin.

"When the only tool you have is a hammer, a sore thumb you will have."

feline_dracoform14-Apr-04 6:38

14-Apr-04 6:38

ah, thanks Smile | :)

i hadn't discovered that yet.

based on a quick experiment, this detects empty tags. when i get onto more complicated cases this is likely to be a problem.

specifically, i have a column "Collected" which is produced by merging together upto three different dates.

i had planned to lay this out as:

<collected>
  <start>
    <date>1/5/2004</date>
    <time>11:52</time>
  </start>
  <end>1/5/2004
    <date>1/5/2004</date>
    <time>12:16</time>
  </end>1/5/2004
</collected>

unless i go through and remove multiple levels of empty tags, i am likely to end up with:

<collected>
  <start/>
  <end/>
</collected>

which is going to match to //collected even though there is no data in it.

is there any way to detect that there is no data in this collected block? i am trying to avoid making assumptions about empty tags always being removed, since this might cause me a problem when i have to insert non breaking spaces into some of my table cells.

Philip Fitzsimons14-Apr-04 22:45

14-Apr-04 22:45

if test="string-length(//collected/*/text()) > 0"

the * means any child node of collected..

"When the only tool you have is a hammer, a sore thumb you will have."

feline_dracoform15-Apr-04 0:19