Trying to bring in a table of static data

Questions about Thunkable AI
1

I’m at wits end (and bottom of the tokens bucket) trying to bring tables of static data into an A.I. app.

The bot says I can:
"Export your Google Sheet or CSV as JSON (use a free converter like csvjson.com)

Save it as a file in your project assets (e.g., foods-data.json)

Import and seed on first app load:"
But there is no capacity to put tabular data into the project assets.

I went and set up an account at Supabase, only to find that this setup is only for all-cloud storage projects. A key feature of my application is that all data is to be stored locally.

I did manage to get the bot to read a CSV file from a shared Google Sheet, but then one thing or another goes haywire…

The bot (un)helpfully started this app with a small hard-coded table of generic data, 20 rows of six columns. I now want to replace that table directly with a table of 5280 rows of 52 columns. All this data is fixed and permanent, the user will be able to view an item in detail and copy it to a local user table, with edits as desired. None of those forms have been built yet.

The bot simply fails in a wild variety of ways, from scrambling the new columns to losing calls to the original table name to variable undefined errors. I tried asking it import a new table. I tried over-writing the old table. I asked it to completely delete to old table and start new.

I tried paring down the new data to 100 rows to speed things up, but that didn’t help much.

How can I get this done?? Again, I just want a pile of static data, permanently in the application. I’m not worried about updates outside of the user installing a new app version.

EDIT: I peeked at the code before and after an attempt at importing the 100 line table. I asked it to directly replace FOOD_CATALOG…

err-pre_import_100
err-pre_import_1001490×387 101 KB

err-post_import_100
err-post_import_1001486×521 201 KB

First of all it made a new variable name, FOOD_CSV_DATA, instead of reusing the existing name. Also, this looks like one very long invalid line of code, nothing like the old multi-line list. Does this have to do with the displayed pop-up error?

And I bet there are still calls for FOOD_CATALOG.

2

As of right now, having a local table of table of 5280 rows of 52 columns is something that the platform is not designed to do at the moment. It might be possible so this is what I would try if I want to give it the best chance of working. All of this would be done in discuss mode.

First, I would tell the AI to get rid of Food CSV DATA.
Then I would tell the ai to replace the FOOD_CATALOG data with these new 5 rows (then paste in 5 rows of food data).
Then test to see if the app works as you intend with those 5 rows. Then I would start to prompt the AI to add in more and more rows. I would again start with adding 5 more, then 10, then 20. Start small and if it proves to work, then keep increasing the amount.