Closed RCura closed 6 years ago
This is actually expected behaviour by dplyr, these random characters are names for the subquery result. Could you try to make a reproducible example of queries that don't work any more because that should not happen.
Can't reproduce the bug: :/ I changed a lot of things considering the way I call my queries, so, must have been a bad combination of DBI/dbplyr/MonetDBLite.
Especially, you might want to investigate the differences of the way R is connecting to the database.
MonetDBLite::src_monetdblite("path")
and DBI::dbConnect(MonetDBLite::MonetDBLite(), "path")
produce different objects, without the same capacities.
For example, you can't disconnect a connection made through src_monetdblite()
, and thus you need to shutdown the server using MonetDBLite::monetdblite_shutdown()
.
Also, some connectors allow uppercase tables and fields while some others don't and fail/produce a corrupted DB.
The results are also different considering the way to copy data in the db : copy_to
and DBI::dbWriteTable()
.
I don't have the time to investigate later on this right now, but having one good and supported way to use MonetDBLite would for sure be an improvement.
Anyway, I finally succeeded using MonetDBLite as a backend for a shiny app dealing with lots of data, and that's a huge improvement in performance and possibilities over the previously used SQLite. Kudos and cheers !
When adding multiple filters from
dplyr
/R
MonetDBLite adds random characters in the queries :Here, it's not a real problem as the query can still work. But when adding more filters, it becomes really messy :
and in queries that I work on, this doesn't work at all anymore : the query isn't correct anymore and raises errors.
When filtering in multiple attributes in one filter, it works as it should :
In my application, I have to rely on multiple filters put on multiple objects, and can't write the whole filter at once. So, this is a really big problem for my app :/