Q Research Software

User wanted to make a new variable with categories based on responses to a few questions. They needed to use JavaScript to do this, but it would be cool if we could somehow expand on our filtering tool or create something similar where they could use dropdowns and selections to build logic for each new category So instead of using code like this: !image-20230310-183455.png|width=645,height=225! Have a dropdown at the top of the filtering tool to add/Edit a new category and building the logic using the built-in menus: !image-20230310-184014.png|width=1359,height=673! We’d still want a preview window of the responses vs results because that’s SO helpful when confirming the logic is sound and covers all respondents' responses) like: !image-20230310-183439.png|width=653,height=303!

A user has suggested that we allow the full attribute scale or values to appear in new variables created using the Filter One Question by Another Question Script. In his specific case, he applied the filter script, but most of his new variables that were created did not include the 1-10 scale that was present in the original question since there weren't respondents who had selected some of the scale ratings. He then had to apply a separate, custom QScript to force the 1-10 point scale to appear for all filtered variables. Example of filtered variable scale: !image-2021-07-13-10-09-50-080.png|thumbnail! And full scale from original, unfiltered variable: !image-2021-07-13-10-10-17-318.png|thumbnail!

User writes: “I agree, I would expect that adding a different filter to the first position would override whatever's already in that position (e.g. the 'answered no to question x' in my example). The part that is unexpected as a user is that the filter dropdown menu behaves sometimes like the position of filters is important and sometimes like it isn't. For example, changing dropdown 2 when there is already a filter 1 will change filter 2, but changing dropdown 2 when there isn't a filter 1 will change filter 1. If I use dropdown 2 twice in a row, I can get a different result (changing filter 1 then changing filter 2), however if I used the first dropdown twice in a row, it would always change filter 1. In computer speak I would expect this operation to be idempotent i.e. doing something in dropdown 2 would always change filter position 2.“ Ultimately, the user us using for a different user interface to remove and add filters to multiple tables.

A couple things here for the Filter One Question by Another automation: If you have NETs created in the data reduction for the _filter by_ question, they will be used to create variables in the new filtered question. However, if you have nets in the question to be filtered, variables for those are not created. You must recreate all the nets in the new filtered question. The overall/main NET in the _filter by_ question does not have a variable created for it using the automation (just the sub-nets do). Sometimes people use this as their total row/column and rename it to something more custom and would like to create a variable for it using the automation as well. You can’t really do this if your _filter by_ question is a Pick One as the NET will be NA in the filtered question as the base is different for each category. You’d have to create some custom JS or R variable to do the NET calculation and merge it into the larger question.

Users with dashboards in different languages may want to modify the text of the “Clear All Filters” option in the Page Filter dropdown in View Mode. In specific case, wanted to change it to “Nat Rep filter”.

Filters work a bit different in Displayr to Q but they should work the same to be consistent.  In Q, you can apply lots of filters and they'll be "AND"-ed together.  In Displayr, filters that are part of a Pick Any question (or a Binary - Multi ) will be "OR"-ed together. To fix it, you will need to create filters in Displayr that are not part of the same variable set but treat the inputs from multiple variables as an OR statement.  For example, if your filter in Displayr is for users of, say, brands, then you may need to create a single combined variable that holds whether any of the relevant brands were selected or not. Alternatively, you can click No in Q, which will still treat the filters as present, even though you won't be able to see them in Q.

The function Filter One Question by Another returns 1’s and 0’s throughout, and ignores missing values in the question being filtered. Thus, when the filtered variable is returned cases that had missing data in the original variable have 0’s for not selected, but no missing values. The feature request is to add an option to respect missing values in the filtered question, or to ignore them and return as currently.

The user is happy with this custom rule: [ https://help.qresearchsoftware.com/hc/en-us/articles/4412661834383-How-to-Hide-Rows-Below-a-Specified-Value|https://help.qresearchsoftware.com/hc/en-us/articles/4412661834383-How-to-Hide-Rows-Below-a-Specified-Value|smart-link ] ([ https://help.qresearchsoftware.com/hc/en-us/articles/4412661834383-How-to-Hide-Rows-Below-a-Specified-Value|https://help.qresearchsoftware.com/hc/en-us/articles/4412661834383-How-to-Hide-Rows-Below-a-Specified-Value|smart-link ] ) However, he also then wants to be able to filter his table out of the records who are now shown. As an example, he wants to show those that have a value higher than 50%, but those under 50% should be removed from the base.