spark_apply error specifying column names

Question

I am running sparklyr in local mode from RStudio in Windows 10:

spark_version <- "2.1.0"
sc <- spark_connect(master = "local", version = spark_version)
df <- data.frame(id = c(1, 1, 2, 2), county_code = c(1, 20, 321, 2))
sprintf("%03d",as.numeric(df$county_code))

df_tbl = copy_to(sc,df, "df_tbl", overwrite = TRUE)
df_tbl %>% summarise(sum = sum(county_code)) %>% collect() ## this works

## this does not:
df_tbl %>% 
   spark_apply(function(e) data.frame(sprintf("%03d",as.numeric(e$county_code), e),
                                     names = c('county_code_fips', colnames(e))))

The last line returns the following error:

Error in file(con, "r") : cannot open the connection
In addition: Warning message:
In file(con, "r") :
  cannot open file 'C:\Users\janni\AppData\Local\Temp\RtmpELRVxu\file4ab817055ccc_spark.log': Permission denied

This happens on both my laptop and desktop. I tried running RStudio as an administrator, but it would not change anything.

Answer 1

It seems the problem comes from the way names specified for the spark_apply.

One option is that you can do without this without the names argument.

df_tbl %>% 
 spark_apply(function(e) data.frame(county_code_fips = 
                                 sprintf("%03d",as.numeric(e$county_code)), e))

## Source: spark<?> [?? x 3]
#  county_code_fips    id county_code
#  <chr>            <dbl>       <dbl>
#1 001                  1           1
#2 020                  1          20
#3 321                  2         321
#4 002                  2           2

Since names does not have access to the e in the function inside the spark_apply, you have to use the names from the tbl.

df_tbl %>% 
 spark_apply(function(e) 
     data.frame(sprintf("%03d",as.numeric(e$county_code)), e), 
     names = c('county_code_fips', colnames(df_tbl)))

## Source: spark<?> [?? x 3]
#  county_code_fips    id county_code
#  <chr>            <dbl>       <dbl>
#1 001                  1           1
#2 020                  1          20
#3 321                  2         321
#4 002                  2           2