You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@arrow.apache.org by "Mauricio 'Pachá' Vargas Sepúlveda (Jira)" <ji...@apache.org> on 2022/02/27 22:18:00 UTC

[jira] [Created] (ARROW-15792) Error when reading integer-type columns

Mauricio 'Pachá' Vargas Sepúlveda created ARROW-15792:
---------------------------------------------------------

             Summary: Error when reading integer-type columns
                 Key: ARROW-15792
                 URL: https://issues.apache.org/jira/browse/ARROW-15792
             Project: Apache Arrow
          Issue Type: Bug
          Components: R
    Affects Versions: 7.0.0
            Reporter: Mauricio 'Pachá' Vargas Sepúlveda
             Fix For: 6.0.1
         Attachments: arrow-error.zip

I was reading files that I created on the same laptop, same R version and Arrow version 7.0.0. The error I get when reading int32-type columns is
{code:java}
> source("~/github/un_escap/hs12-historic-series/00-create-historic-series.R")
2012
Error in `handle_csv_read_error()`:
! NotImplemented: Function 'equal' has no kernel matching input types (array[int32], scalar[string])
Run `rlang::last_error()` to see where the error occurred. {code}
Here's a minimal example of the code that caused this
{code:java}
library(arrow)
library(dplyr)

d <- open_dataset("aggregate_level=0",
                  partitioning = c("trade_flow", "year", "reporter_iso"))

d %>%
  select(reporter, reporter_code, reporter_iso) %>%
  collect() {code}
After going back to Arrow R package v6.0.2, it worked. Here "reporter_code" is numeric.

I've attached a zip with the data.

 



--
This message was sent by Atlassian Jira
(v8.20.1#820001)