XML: XMLMAP NAMESPACES resulting in 0 observations

[KevinViel]

If I omit the NAMESPACES lines, then my table has the expected number of observations.  If I include it, then I get zero observations:

<?xml version="1.0" encoding="windows-1252"?>

<SXLEMAP name="define_2a" version="2.1">

  <NAMESPACES count="1">

  <NS id="1" prefix="def">http://www.cdisc.org/ns/def/v2.0</NS>

  </NAMESPACES>

  <TABLE name="ItemGroupDef">

ItemGroupDef does not reference the NS id = "1".  If any sees the issue or has a hint, then I would be happy to read it.

Thank you,

Kevin

[Shmuel]

What is the connection to sas ?

What is your sas code?

[KevinViel]

filename MAP

         "O:\krviel\cdisc-definexml-2.0.0-1.6\referencexml\define_2a.map"

         ;

libname define

        xmlv2

        "O:\krviel\cdisc-definexml-2.0.0-1.6\define.xml"

        xmlmap = MAP

        ;

data ItemGroupDef ;

  set define.ItemGroupDef ;

run ;

filename MAP

         clear

         ;

libname define

        clear

        ;

I am user v9.4 on Windows, but I started in with the Unicode Support.

Thank you,

Kevin

[KevinViel · Accepted solution]

The issue is that I omitted namespaces.  Chevell (Parker) of SAS Tech Support, whom I thank, suggested that I debug by creating the automap.  The namespaces should be:

<NAMESPACES count="3">

  <NS id="1" prefix="xlink">http://www.w3.org/1999/xlink</NS>

  <NS id="2" prefix="def">http://www.cdisc.org/ns/def/v2.0</NS>

  <NS id="3" prefix="">http://www.cdisc.org/ns/odm/v1.3</NS>

</NAMESPACES>

Then the paths would be of this form:

<PATH syntax="XPathENR">/{3}ODM/{3}Study/{3}MetaDataVersion/{3}ItemGroupDef/{3}ItemRef/{1}WhereClauseRef</PATH>

I now have the observations:

12709 data ValueListDef ;

12710 set define.ValueListDef ;

12711 run ;

NOTE: There were 3 observations read from the data set DEFINE.ValueListDef.

NOTE: The data set WORK.VALUELISTDEF has 3 observations and 3 variables.

NOTE: DATA statement used (Total process time):

real time 0.01 seconds

cpu time 0.00 seconds

It should have been a clue that the resulting data sets had 0 observations, but the expected number of variables.

HTH,

Kevin

PS Note that between the OP and this post that ID for "def" changed.