JavOrraca commented on issue #2508:
URL: https://github.com/apache/arrow-adbc/issues/2508#issuecomment-2652102802
Still digging into my issues at work and I appreciate your feedback. Below
are some new updates for you:
- Ending my `read_adbc()` query with `LIMIT xxx` didn't impact performance
vs using `SELECT TOP xxx`
- Using ADBC, I'm now unable to complete my profiling with R's `profvis`, or
benchmarking with `bench`, on the biggest data set in my tests (1 million rows
x 75 columns)
- My RStudio session crashes with no detailed logs/notes, and this is
obviously odd since last week I ran the tests in the picture above 🤷♂️
- I tried the same process on several different servers / containers and
experienced the same issue even after re-installing Go, `adbcdrivermanager`,
and `adbcsnowflake`
- More background about my working environment: I'm using a cloud platform
developed in-house that launches a user namespaced container running Ubuntu 22,
and it's from this container that I launch R and RStudio
- This is a mature platform at my company that we've had for 7+ years and
despite some minor inconveniences, it works really well for our data analysis
and modeling needs including Python, R, git, etc.
- I checked Dynatrace and nothing sticks out in terms of exceeding allowed
CPU or RAM usage, and network analysis (traffic inbound and NIC packets
received) looks "normal" comparing my ODBC vs ADBC approaches
And one final question: Is there an option in `adbc_database_init` or
`adbc_connection_init` that I can set for the high precision option,
potentially something like the pseudo code below? Thanks here... Tried Googling
this and it's not clear what I should tweak in the driver options. 🙏
```{r}
library(adbcdrivermanager)
db <- adbc_database_init(
adbcsnowflake::adbcsnowflake(),
# connection string details
high_precision = TRUE
)
conn_adbc <- adbc_connection_init(db)
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
