There's a really fine line between needing a spreadsheet and needing a database and I've not yet found it. It's probably more fuzzy than I realized but I have participated on so many programming projects that amounted to a spreadsheet that lived too long.
this post was submitted on 27 Sep 2023
426 points (96.7% liked)
Programmer Humor
32371 readers
562 users here now
Post funny things about programming here! (Or just rant about your favourite programming language.)
Rules:
- Posts must be relevant to programming, programmers, or computer science.
- No NSFW content.
- Jokes must be in good taste. No hate speech, bigotry, etc.
founded 5 years ago
MODERATORS
Does it need to be accessed by multiple people? Does it need to be updated frequently? Does it need to be accessed programmatically? Does performance matter? If you answered "yes" to any of these questions, you should probably use a database.
If it's something you interact with manually, has less than 100,000 rows, and is mostly static, then sure, use a spreadsheet.
I used to have some scripts to convert and merge between CSV and sqlite3. Even a lightweight db like sqlite3 has value. Google Sheets fills some of the gaps with its QUERY statement but I still find it a bit awkward to use a lot of the time.
Google sheets works just fine for accessed by multiple people.
The line is probably somewhere on machine vs human readable.
Performance. If you get 30k transactions per second don’t even SAY spreadsheet lol
Per second? If you get that many per day, I wouldn't touch it with a spreadsheet.
I can answer yes to all of these questions but still use a spreadsheet. I understand your point, but I feel even with these the line is still gray.
I just checked and my largest spreadsheet currently has 14,300 lines across 12 tabs. Each tab contains the same information just pulled from a separate form. So each tab is linked to a form to update automatically when someone submits a new response. We then process these responses periodically throughout the day. Finished responses are color coded so a ton of formatting. Also 7+ people interacting with it daily.
Then we have a data team that aggregates the information weekly through a script that sends them a .csv with the aggregate data.
The spreadsheet (and subsequent forms) are updated twice a year. It was updated in June and will be updated again in December. It’s at 14k now and will continue to grow. We’ve been doing this through a few iterations and do not run into performance issues.
At some point you end up surpassing databases and end up with a giant pile of spreadsheets called a data warehouse
As soon as you stop data maintenance per hand, start using a db.
The article doesn't seem to say what type of database they moved to, I'd like to imagine it's an excel spreadsheet...
The problem with the spreadsheet was rate limiting by Google. I like to imagine the have the spreadsheet copy and pasted. Then split the requests to two different spreadsheets, doubling the amount of requests they can do.
Dear god, parallelized excel spreadsheet databases!
SAP S/4 HANA is not mental illness
It's worse
Your physical health and everyone you love will suffer too
Sandhur Anziege Programm, or "Hourglass Displaying Program" in English, first started out as a hardware stress tester. It only made sense that it would evolve into a human stress testing program from there.
I don't wanna give them bad ideas, but the only logical next step is to have 2TB of CPU cache.
All the cool kids use Microsoft Access
Not going to lie. I have made it really far using Google sheets as a database until I had to move to a full DB. Google Sheets is just an interface to a Google run DB anyways. A lot of the time I would just load sheets(tables) into data structures from a SQL library and use it like a regular database.
I use Google sheets on my personal projects.
- tracking expenses
- tracking my ride logs from my ev
Google app script pickup email received from banks and update the transaction automatically. I modify it using web based form of needed.
The ev company has api. It will be called on periodic basis to get the ride details and update the sheets. I use telegram api to interact, by triggering the api webhook, getting charts, etc.,
Initially I setup OCR to extract the ride information, so the process was like
- send an image that has ride details to a telegram bot
- the bot saves the image in Google drive
- the image would be opened in docs
- the text would get extracted and stored in Google sheets as data
- edits can be done using telegram webapp
- sheets provide charts for analysis
- built a process using sheets to handle telegram bot for maintaining users, approvals, etc.,
It would be hard if I have used any other services.. setting up OCR preparing charts etc.,
thank god postgres is still safe!
Nobody is dumb enough to insult postgres - we'll fucking burn you at a stake for heresy like that.