Regression: reading multiple Parquet files

[krlmlr]

Introduced in c52bef5.

if (!file.exists("1.parquet")) {
  duckplyr::compute_parquet(data.frame(a = 1, b = 2), "1.parquet")
}
#> Loading duckplyr
#> # A duckplyr data frame: 2 variables
#>       a     b
#>   <dbl> <dbl>
#> 1     1     2
if (!file.exists("2.parquet")) {
  duckplyr::compute_parquet(data.frame(a = 2, b = 3), "2.parquet")
}
#> # A duckplyr data frame: 2 variables
#>       a     b
#>   <dbl> <dbl>
#> 1     2     3

urls <- c("1.parquet", "2.parquet")

con <- DBI::dbConnect(duckdb::duckdb())
duckdb:::rel_from_table_function(con, "read_parquet", list(list(urls)))
#> Error:
#> ! {"exception_type":"Invalid Input","exception_message":"Failed to cast value: Unimplemented type for cast (VARCHAR -> r_string)"}

^{Created on 2025-01-20 with reprex v2.1.1}

duckdb 1.1.3-1

if (!file.exists("1.parquet")) {
  duckplyr::compute_parquet(data.frame(a = 1, b = 2), "1.parquet")
}
#> Loading duckplyr
#> # A duckplyr data frame: 2 variables
#>       a     b
#>   <dbl> <dbl>
#> 1     1     2
if (!file.exists("2.parquet")) {
  duckplyr::compute_parquet(data.frame(a = 2, b = 3), "2.parquet")
}
#> # A duckplyr data frame: 2 variables
#>       a     b
#>   <dbl> <dbl>
#> 1     2     3

urls <- c("1.parquet", "2.parquet")

con <- DBI::dbConnect(duckdb::duckdb())
duckdb:::rel_from_table_function(con, "read_parquet", list(list(urls)))
#> DuckDB Relation: 
#> ---------------------
#> --- Relation Tree ---
#> ---------------------
#> read_parquet([1.parquet, 2.parquet])
#> 
#> ---------------------
#> -- Result Columns  --
#> ---------------------
#> - a (DOUBLE)
#> - b (DOUBLE)

^{Created on 2025-01-20 with reprex v2.1.1}

[krlmlr]

Specifically, 721d611 is the trigger -- motivated by the upstream change, but not caused by it. Interesting.

[github-actions]

This old thread has been automatically locked. If you think you have found something related to this, please open a new issue and link to this old issue if necessary.