The company I work for has as core activity managing geographic information and has a lot of datasets delivered. We explore to apply data profiling techniques and are evaluating Talend DQ (data profiler).

Data mining/data profiling is OK for Nominal, Interval, text but not for SPATIAL data (perhaps an idea for OTHER...).
Quite a bunch of nice statistics can be distilled from geometries.

So I've tried the User defined indicators on SQL (with MS SQL-Server). Indicators can be quit easily created for generating statistics. But here comes the 'issue':

When defining a 'user defined query' as for example:

> 'SELECT <%=__COLUMN_NAMES__%>.STSrid, COUNT(<%=__COLUMN_NAMES__%>.STSrid )
FROM <%=__TABLE_NAME__%>
WHERE <%=__COLUMN_NAMES__%>.STIsValid() = 1
<%=__AND_WHERE_CLAUSE__%>
GROUP BY <%=__COLUMN_NAMES__%>.STSrid'

with > Indicator category = 'Frequency analysis'

a. The <analysis results> do not mention what indicator is used (only mentioning user defined indicator) so results are not readable, as no one knows what the graphic is all about, neither the analysis results.

b. In the graphics or Analysis result, No column-label is displayed.
I'd like to have the columnnames (eg GEOM.STSrid) or a way to define aliases as for example:

'SELECT
<%=__COLUMN_NAMES__%>.STSrid as 'STSrid'
, COUNT(<%=__COLUMN_NAMES__%>.STSrid ) as 'Count'

FROM <%=__TABLE_NAME__%>
WHERE <%=__COLUMN_NAMES__%>.STIsValid() = 1
<%=__AND_WHERE_CLAUSE__%>
GROUP BY <%=__COLUMN_NAMES__%>.STSrid'

Is there another way around easing readibility ?
Please advice.

Bertrand Ruzette