Request to Disable Stale Data Feature Due to Ongoing Bugs

Ludovic V. · 2024-10-30 20:03:52.571592+00

Hello, can you give users the chance to deactivate the stale data feature please? I've have dozens of bugs for months with it that already made me spent tens of thousands of credits for nothing. It happened again just now on a thousands of lines table. Feels like I'm getting forced to waste my credits and the fact that you're using the conditional form "it may have" makes me even less confident about the benefit of this feature for your users. I would rather pay more per credit than be forced to waste them. :slightly_smiling_face:

Request to Disable Stale Data Feature Due to Ongoing Bugs | Clay

6 comments

· Sorted by Oldest

Ludovic V.
·
·
Got a part of an answer from your video about stale data: The fact that when you change a column and ask to only run it for cells that haven't run or have errors, and that it'll automatically add everything else as stale data is super confusing, in a very positive way for Clay but negative one for users.. In your video:
C1 and D2 haven't run? No
C1 and D2 have errors? No
So for anyone who haven't watch that specific video, nothing should happen to those cells based on the command performed right? Yet they get impacted big time. That's one way to loose a lot of credits. I've made a very subtle change to a column, "improving it" but I never wanted to waste re-run the thousands of lines that used the almost similar one before. Not possible says Clay. Surely you can find a way to make it understandable before it happens 🙂
Loading attachments
Ludovic V.
·
·
This one is still unknown. I have a URL which is manual data. It looks up for the person on LinkedIn. Then i ask ChatGPT to figure out if the person speaks French or English. Got many lines in stale data without having edited the column at all. Stale data for a language? 100% of the time it'll be the same, so it's an even more waste. Would love to know the reasons behind (might need to make a bug ticket for this one, since the first one seems tricky but expected)
Loading attachments
Ludovic V.
·
·
Let me know! Still think giving users the chance of activating it or deactivating it will help. Not everyone needs/wants/can afford to spend thousands of credits to update the entire column where a subtle change was made 🙂
Ludovic V.
·
·
Last one: Why is there no way to filter stale data? in the bug above with the French/English ChatGPT scenario where I changed nothing and I have this stale data thing showing up, to limit the waste, I tried to filter the stale data and force run only those rows, but it seems it's not possible. Does it mean right now when you have stale data, it it's either you do it manually one by one or if it's too big of a job, you have to force run all the rows, included those that didn't have stale data? Yet again waste of credits. This update with stale data really feels different than all Clay release: Unclear, confusing, limited in filtering options. 😵‍💫
Loading attachments
Arturo O.
·
·
Hey Ludovic, thanks for the message and feedback! I've share all of this with our engineering team to consider additional options. I'd like to emphasize, as shown in the example, that this is not exclusive to Clay but that's the definition when it comes to APIs overall because API endpoints have a unique structure (URL) when they're processed, changing its values (even outside of Clay) that would be considered an entirely new request (new URL), unless nothing in its structure changed. Of course, I get it that there may be use cases where there's a need to update stuff because edge cases come up thus could cause a downstream effect for the dependencies resulting in this outcome. We do have some things on the roadmap that may help and will attach your feedback to that internal topic. Finally, happy to help you out, how many credits did you go through here?
Channeled
APP
·
·
Hi Ludovic V.! This thread was recently closed by our Support team. If you have a moment, please share your feedback:

Ludovic V.
·
·
Got a part of an answer from your video about stale data: The fact that when you change a column and ask to only run it for cells that haven't run or have errors, and that it'll automatically add everything else as stale data is super confusing, in a very positive way for Clay but negative one for users.. In your video:
C1 and D2 haven't run? No
C1 and D2 have errors? No
So for anyone who haven't watch that specific video, nothing should happen to those cells based on the command performed right? Yet they get impacted big time. That's one way to loose a lot of credits. I've made a very subtle change to a column, "improving it" but I never wanted to waste re-run the thousands of lines that used the almost similar one before. Not possible says Clay. Surely you can find a way to make it understandable before it happens 🙂
Loading attachments
Ludovic V.
·
·
This one is still unknown. I have a URL which is manual data. It looks up for the person on LinkedIn. Then i ask ChatGPT to figure out if the person speaks French or English. Got many lines in stale data without having edited the column at all. Stale data for a language? 100% of the time it'll be the same, so it's an even more waste. Would love to know the reasons behind (might need to make a bug ticket for this one, since the first one seems tricky but expected)
Loading attachments
Ludovic V.
·
·
Let me know! Still think giving users the chance of activating it or deactivating it will help. Not everyone needs/wants/can afford to spend thousands of credits to update the entire column where a subtle change was made 🙂
Ludovic V.
·
·
Last one: Why is there no way to filter stale data? in the bug above with the French/English ChatGPT scenario where I changed nothing and I have this stale data thing showing up, to limit the waste, I tried to filter the stale data and force run only those rows, but it seems it's not possible. Does it mean right now when you have stale data, it it's either you do it manually one by one or if it's too big of a job, you have to force run all the rows, included those that didn't have stale data? Yet again waste of credits. This update with stale data really feels different than all Clay release: Unclear, confusing, limited in filtering options. 😵‍💫
Loading attachments
Arturo O.
·
·
Hey Ludovic, thanks for the message and feedback! I've share all of this with our engineering team to consider additional options. I'd like to emphasize, as shown in the example, that this is not exclusive to Clay but that's the definition when it comes to APIs overall because API endpoints have a unique structure (URL) when they're processed, changing its values (even outside of Clay) that would be considered an entirely new request (new URL), unless nothing in its structure changed. Of course, I get it that there may be use cases where there's a need to update stuff because edge cases come up thus could cause a downstream effect for the dependencies resulting in this outcome. We do have some things on the roadmap that may help and will attach your feedback to that internal topic. Finally, happy to help you out, how many credits did you go through here?
Channeled
APP
·
·
Hi Ludovic V.! This thread was recently closed by our Support team. If you have a moment, please share your feedback: