Forums: Define.xml
This message is "Variable Data Types in the dataset must match the variable data types described in the data definition document (define.xml)"
Can you provide the name and datatype of STUDYID listed in both your define.xml and SAS Dataset.
I can send you screen captures in a Word document, can't seem to attach to this forum post.
[ I am: carmichael_oneill@yahoo.com ]
But anyway:
In define.xml, I see in dataset PG:
Variable: STUDYID
Label: Study Identifier
Type: text
Controlled Terms or Format: <blank>
Origin: Protocol
Role: Identifier
In SAS xpt contents:
Variable: STUDYID
Type: Character
Length: 20
Format: <blank>
Informat: <blank>
Label: Study Identifier
Perhaps
define.xml STUDYID type = "text" <>
SAS xpt STUDYID type="Character"
As a quick test, if you change your define.xml to be type Character, does the warning go away ?
i see what's pasted below - not sure where to change the STUDYID attribute from "text" to "Character"
?
<ItemGroupDef OID="PG" Name="PG" Repeating="No" IsReferenceData="Yes" Purpose="Tabulation" def:Label="Pharmacogenomics" def:Structure="One record per lab test per time point per visit per subject" def:DomainKeys="STUDYID USUBJID PGCAT PGTESTCD PGDTC VISITNUM" def:Class="Findings" def:ArchiveLocationID="Location.PG"> <ItemRef ItemOID="STUDYID" OrderNumber="1" Mandatory="Yes" Role="Identifier" RoleCodeListOID="RoleCodeList"></ItemRef> etc.
I do see this:
<ItemDef OID="STUDYID" Name="STUDYID" DataType="text" Length="20" Origin="Protocol" Comment="Value: XXX-YYY." def:Label="Study Identifier"> It looks like it's defined once, which makes sense. I don't see other DataTypes as "Character", though - and it seems to only be flagging errrors in PG and SUPPxx. I will continue to look at the define.xml text and see if there's something odd there.
I think it's a define.xml thing, I'll have to check what our vendor did. SD0059 is being flagged for the last set of domains in the define, PG then all SUPPxx. Nothing else before then, and all S0059 after then.
I will run the OpenCDISC validation on the define.xml, maybe that will show something.
Thanks for the help, I'll consider my original post closed.
If anyone has any idea why this is happening for one SDTM dataset and all SUPPQUALs ... maybe something to do with define.xml construction? It's not clear why I'm getting this error. In this example, it looks like STUDYID within domain PG is identified in a way consistent with what it is.
(Using Version 1.4)
(see attached pdf document which features:
OpenCDISC screencap
define.xml screencap
SAS Universal Viewer screencap
)
(file not attaching, can't seem to share the image - post to this thread if you like, and I can reply personally for the example in email)