Find partial (but closes) matches for terms in 2 differents tables

Would it be possible to use Baserow to find partial (but closes) matches for terms in 2 differents tables (specific cells) and show the associated lines data?
Ex.

In table 1, in the choosen column for this query, there is a cell with ‘behind the word mountains’
In table 2, in the choosen column there is a cell with ‘behind words mountains’

I would like to output both cells thats match, along all data on the same line for both tables.

Hey there :slight_smile:

This could be achieved by using Link row and lookup fields.

To also search data from another table you would need to add that data to your current table by first making a link row field connected to that table and then you could add a lookup field for every relevant field you want to have “duplicated” in your current table from the other table.

Just being able to run search and it will automatically search multiple tables and merge the results, is not something you can just do by enabling a setting for example.

Hope that helps :slight_smile:

Thanks Alex. I will play with the Lookup field!

But by looking at the doc, I can’t find a parameter to adjust the precision of the search. I assume it have to be exact match?

If it is, it’s not what I’m looking for. It have to be close but not perfect.

Example :

Chateau du Lac de la Tour

would have to match with

chateau lac de tour

Yes that wouldn’t work.

The search does a contains comparison on the whole string that you insert into the search.

What that means is, it will take the search text that you put in, and see if any of the fields contain that search text exactly how you have written it. It won’t re-arrange the letters or words and look for matches based on that.

Any idea how to do it?

It’s probably not possible. I can’t think of a way you can achieve this with formula fields, and our search is non-configurable. @nigel can you confirm?

Really what we need here is the ability to search and get results backed ordered by some “search distance” metric. Right now our search functionality is very exact and basic.

This issue is tracking adding exactly this feature: Results of a contains query should be pre-sorted by levenshtein distance (#547) · Issues · Bram Wiepjes / baserow · GitLab however we haven’t yet figured out the UX of it. Should it be an option in the search box to instead “sort results by closeness to search term” ? Should it instead be a new type of filter you can add to fields?

1 Like

@randomuseranonymous, please like the issue in GitLab if this is important to you. :slight_smile: It helps us with prioritizing features.

2 Likes

As an inspiration, there is some fuzzy search plugin and formulas used in Airtable and Google Spreadsheet