Does data need to be re-imported every time it changes?

[truenorth]

I'm confused about how/when data is available to IA.

Initially data is imported into IA. Is that just a snapshot? When data changes at source, i.e., where it was imported from, does it have to be re-imported? Or is data always live after the initial import?

Thanks in advance.

[qt_ky]

Initially metadata is imported into IA. After that you can run column analysis or other types of analysis on the actual data. The analysis results are stored and available for further review, like a snapshot.

If the source metadata (table structure) changes, my experience has been to import the new metadata and lose the previous analysis results. I'm sure someone knows of a better way.

As the source data itself changes, you may want to periodically submit IA tasks to run various types of analysis again, to measure if your data quality or anomaly issues are improving over time.

[ray.wurlod]

Data are never imported into IA. Only metadata.

Analysis results are stored in tables in the IA database (frequency distributions and the like) and in the metadata repository (XMETA), but never source data.

Analysis results typically would change given changed source data.

You can set baseline and checkpoints to monitor these changes over time, but it does involve running additional analyses.

If the data are coming from a different source, then your IA project may need to have that new source configured (as a Data Connection or as an ODBC data source).

[truenorth]

Thank you, all!