Using Google’s sentiment analysis step… but getting an error
Please see attached screenshot… not sure how to proceed.
Hey @andrew can you please help me here? or point me in a direction where I potentially could fix this?
Hey @Varun_Goel – I’m not super familiar with this step, but I do know it has limits of how much data you can pass into it. How many rows of data are you sending to that step?
Thanks for the quick response!
at first around 48k, then it gave me an error
So then i reduced it under 1000 rows, and I got the error above.
Who can help me with this?
Hey @Varun_Goel – someone else on the team will take a look today. My hunch is that you reached your allocation (I believe we limit it to 1000 rows for a certain time period, maybe 24 hours). @brian or @sachi will have a look though!
1 Like
Ok, let me give it a try in a few…
1 Like
It gave me the same error even with less than 1000 records… @brian
I don’t know the exact way in which the limit is imposed, but I assume it has some sort of time component, so perhaps letting it sit for a little and try again would be the best result - but to that point, the limit of 1000 rows will still be applied.
1 Like
I did leave it sit for a few days and try and it didn’t work… Is it by any chance looking at the 50k records in step 1 of my flow and then crapping out because there are more than 1000 rows ?
Hey Varun - I’ve asked our engineers for some help in figuring out what’s causing the error. I’ll follow up with you once I have more info.
1 Like
I discussed this error with our engineer, and it turns out that there’s unfortunately a known bug that’s causing the error. It has to do with how the step currently handles comments in different languages sent to the API. Basically any non-English data you’re passing through is likely to fail.
We do have the bug listed to fix, but I don’t have a specific timeline for when we expect to be able to address it. I’ll certainly follow up when it’s resolved.
In the meantime, your only workaround is to only pass English comments for analysis, which I know is difficult to manage and not ideal. Apologies for the frustrating experience. Maybe you can find another API to use for Sentiment Analysis and use the API Enrichment step for it.
1 Like
Hey all!
Jumping in here. I do think that non-english languages are supported! You can see the full list here: https://cloud.google.com/natural-language/docs/languages#sentiment_analysis
| Language | ISO-639-1 Code |
|---|---|
| Chinese (Simplified) | zh |
| Chinese (Traditional) | zh-Hant |
| English | en |
| French | fr |
| German | de |
| Italian | it |
| Japanese | ja |
| Korean | ko |
| Portuguese (Brazilian & Continental) | pt |
| Spanish | es |
Are you using a language not on this list?
2 Likes