Parquet date filters
There seems to be an issue with filtering on date columns in parquet files in duckdb-r versions after 1.4.4
I created scripts for Ubuntu 24.04 using p3m.dev package manager to both replicate the issue and demonstrate how it worked in 1.4.4 and posted them in a repo here: [https://github.com/lab1702/duckdb-r-parquet-issue](https://github.com/lab1702/duckdb-r-parquet-issue)
In short, when running the query like this:
library(duckdb)
query <- "FROM 'test_data.parquet' WHERE test_date > today() - INTERVAL 7 DAY;"
con <- dbConnect(duckdb())
df1 <- try(dbGetQuery(con, query))
df2 <- try(dbGetQuery(con, query))
dbDisconnect(con)
print(df1)
print(df2)
On version 1.4.4 you will see two data frames printed, but on later versions only the second data frame has data
and the first one has an error:
````text
Loading required package: DBI
Error in dbSendQuery(conn, statement, ...) :
Binder Error: No function matches the given name and argument types 'age(DATE, INTERVAL)'. You might need to add explicit type casts.
Candidate functions:
age(TIMESTAMP) -> INTERVAL
age(TIMESTAMP, TIMESTAMP) -> INTERVAL
age(TIMESTAMP WITH TIME ZONE) -> INTERVAL
age(TIMESTAMP WITH TIME ZONE, TIMESTAMP WITH TIME ZONE) -> INTERVAL
LINE 1: FROM 'test_data.parquet' WHERE test_date > today() - INTERVAL 7 DAY;
^
ℹ Context: rapi_prepare
ℹ Error type: BINDER
[1] "Error in dbSendQuery(conn, statement, ...) : \n Binder Error: No function matches the given name and argument types 'age(DATE, INTERVAL)'. You might need to add explicit type casts.\n\tCandidate functions:\n\tage(TIMESTAMP) -> INTERVAL\n\tage(TIMESTAMP, TIMESTAMP) -> INTERVAL\n\tage(TIMESTAMP WITH TIME ZONE) -> INTERVAL\n\tage(TIMESTAMP WITH TIME ZONE, TIMESTAMP WITH TIME ZONE) -> INTERVAL\n\n\nLINE 1: FROM 'test_data.parquet' WHERE test_date > today() - INTERVAL 7 DAY;\n ^\n\033[34mℹ\033[39m Context: rapi_prepare\n\033[34mℹ\033[39m Error type: BINDER\n"
attr(,"class")
[1] "try-error"
attr(,"condition")
<error/rlang_error>
Error in `dbSendQuery()`:
! Binder Error: No function matches the given name and argument types 'age(DATE, INTERVAL)'. You might need to add explicit type casts.
Candidate functions:
age(TIMESTAMP) -> INTERVAL
age(TIMESTAMP, TIMESTAMP) -> INTERVAL
age(TIMESTAMP WITH TIME ZONE) -> INTERVAL
age(TIMESTAMP WITH TIME ZONE, TIMESTAMP WITH TIME ZONE) -> INTERVAL
LINE 1: FROM 'test_data.parquet' WHERE test_date > today() - INTERVAL 7 DAY;
^
ℹ Context: rapi_prepare
ℹ Error type: BINDER
---
Backtrace:
▆
1. ├─base::try(dbGetQuery(con, query))
2. │ └─base::tryCatch(...)
3. │ └─base (local) tryCatchList(expr, classes, parentenv, handlers)
4. │ └─base (local) tryCatchOne(expr, names, parentenv, handlers[[1L]])
5. │ └─base (local) doTryCatch(return(expr), name, parentenv, handler)
6. ├─DBI::dbGetQuery(con, query)
7. └─DBI::dbGetQuery(con, query)
8. └─DBI (local) .local(conn, statement, ...)
9. ├─DBI::dbSendQuery(conn, statement, ...)
10. └─duckdb::dbSendQuery(conn, statement, ...)
```
0 条评论