Normalization of Data Types (Tables) vs All-in-one

[Blackbeard]

Good Morning Bubblers,

Currently, I’ve got about 15 data types (I’m using “tables” and “data types” to refer to the same thing here). I have different data types (tables) for donors, the items they’ve donated, donor history, contact history, etc.

Due to reporting issues out of Bubble, and the need for third-party integrations on a transactional level, I’m wondering about the ease of reporting from normalized data tables (separated based on content) vs a monster table (all columns/fields in one table).

From a hierarchy standpoint, data cleanliness, and database efficiency, I understand that normalization is the best practice, but I’m wondering if it’ll cut down on the number of transactions, and merging required to produce an output for my end users.

Thoughts? Examples?

[john3]

I have my data mostly normalized.
I found this gives me the most flexibility in bubble when I want to do something that I didn’t anticipate with the data.

But I don’t think there’s anything wrong with building a giant table either - it just depends on what you are trying to build and the flexibility you want to have for future changes…

[Blackbeard]

Thanks for your reply, John.

The issue I’m facing is that when an end user wants a report that combines two data types, they end up with fields that refer to unique IDs, or they have to do two exports and recombine them.

[john3]

Do you have an example of what you are doing? How does the end user see the unique ID fields?

[Blackbeard]

Hi John,

The end user sees the unique ids when exporting the data to csv, using the “download as csv” workflow action.

This is what the repeating group table data looks like:

Workflow action:

Resulting text:

[john3]

So is your goal to not show the unique id fields on export?

[Blackbeard]

Correct, yes

[john3]

Ah I see - there’s a very deep thread on this and it looks like it can’t be done unfortunately without the bubble team adding the new feature…

Download CSV - pick data

I’m looking for a similar solution – want to create a CSV export that only show specific fields (and needs to exclude some fields in our database too). Any progress on this, either from the Bubble team, or other community members finding creative work-arounds? Thanks!

But I wonder if there’s a way to export your data via an api and let that handle the download aspect so you can pick and choose what fields get shown.

[Blackbeard]

Thanks John. I’ll look through the thread.
There’s another thread I’d created on reporting data out of bubble, too

Reporting Data Out of Bubble?

I’m not terribly enamored with the Export to CSV functionality, as it’s unformatted, and there seems to be some issues with the data it exports (I’ve seen records split between lines, even when the data separator character isn’t present). I’m wondering: how have others sent data from Bubble to another service, to create reports (ex. Google Sheets)? I’ve tried using Blockspring, but that pushes one row at a time, not an entire report’s worth of an export. Bubble doesn’t yet support “walking” a …

[john3]

Ah - I see good to know about the CSV limitations.
I’m sorry there doesn’t seem to be any workarounds.
Best we can do is put it in as a feature request.

[NigelG]

We created a table “on the fly” for reporting, and let the user download that.

It only had the columns they wanted.

There is a thread about it somewhere. Seemed to work quite well.

[uproute]

I’m working through a quick hack of this using Google Sheets + Zapier, but I feel like this is the type of thing that should be or is on the list of things to be accomplished natively with some “select fields for export” concept.

In the interim, the basic idea would be to combine specific data onto a sheet, and have that sheet attached to an email that is triggered to be sent upon the creation of a new spreadsheet. Then the user would receive a “Download Report” Email – so you don’t get the perfect autodownload UX, but still very acceptable UX I think.