Why doesn't my dataset reflect the join conditions I selected? | Experience Community
Skip to main content
Question

Why doesn't my dataset reflect the join conditions I selected?

  • April 30, 2026
  • 3 replies
  • 16 views

Forum|alt.badge.img

I have a dataset created using data modeler. It has multiple join conditions, however, the resulting output dataset does not have the expected data. The fields used as the join conditions have the same field type, and the values do not contain spaces or other hidden characters. In the resulting dataset, it seems that no values from the right side of the join were carried over (they’re all N/A). That includes the survey response's value for distribution id (Q_EMD), which is one of the join conditions so it SHOULD have a value. I have inspected the data sources themselves, and there are rows in each for which all three of the join conditions are true.

Is anyone else having this problem? Have you had this problem and then resolved it? Is there something that I may be missing?

3 replies

Forum|alt.badge.img+3
  • Level 2 ●●
  • April 30, 2026

For the data to be displayed, all values must match. If any join point fails, the entire response remains blank.

My recommendation would be to use only the email-based join. As long as one data point matches, the response will be linked in the dashboard.


Forum|alt.badge.img

All three datapoints match, thus my confusion. I can export my survey responses, and part of the funnel (by using a record table that sits on just the funnel - no joined dataset or anything), then match them up in Excel. I know they match.

Using just email won’t work well in this case, as this is a new dataset on a survey that was also distributed 12 months ago, so it ends up with a few rows that represent last year’s responses but this year’s email distribution data, hence the join condition for distribution ID.


vgayraud
QPN Level 7 ●●●●●●●
Forum|alt.badge.img+64
  • QPN Level 7 ●●●●●●●
  • May 1, 2026

Hi,

Do you have large datasets? Keys with values with more than 50k instances aren’t supported. It could easily be the case with Survey ID or Distribution ID.