ARROW-3818: [R] Table$from_batches

[romainfrancois]

https://issues.apache.org/jira/browse/ARROW-3818?filter=12344983

[romainfrancois]

library(arrow)
#> 
#> Attaching package: 'arrow'
#> The following object is masked from 'package:utils':
#> 
#>     timestamp
#> The following objects are masked from 'package:base':
#> 
#>     array, table

b <- record_batch(tibble::tibble(int = 1:2, dbl = rnorm(2)))
tab <- table_from_batches(b,b,b)
tab
#> arrow::Table
as_tibble(tab)
#> # A tibble: 6 x 2
#>     int     dbl
#>   <int>   <dbl>
#> 1     1 -0.267 
#> 2     2 -0.0270
#> 3     1 -0.267 
#> 4     2 -0.0270
#> 5     1 -0.267 
#> 6     2 -0.0270

# this supports !!! splicing so
batches <- rep(list(b), 20L)
tab <- table_from_batches(!!!batches)
tab
#> arrow::Table
as_tibble(tab)
#> # A tibble: 40 x 2
#>      int     dbl
#>    <int>   <dbl>
#>  1     1 -0.267 
#>  2     2 -0.0270
#>  3     1 -0.267 
#>  4     2 -0.0270
#>  5     1 -0.267 
#>  6     2 -0.0270
#>  7     1 -0.267 
#>  8     2 -0.0270
#>  9     1 -0.267 
#> 10     2 -0.0270
#> # … with 30 more rows

^{Created on 2019-02-05 by the reprex package (v0.2.1.9000)}

[romainfrancois]

I'm just not sure about the naming. In python the py.Table is exposed but in the R package, the classes are not really meant to be used directly, so defining a static method does not really work, as of now the class is arrow:::`arrow::Table` ...

We could expose certain classes a first class citizens of the package if we really want the syntax Table$from_batches()

[romainfrancois]

Actually #3565 is better because it does not need a separate table_from_arrays() function, as the table() itself can handle list of record batches

[xhochy]

Rebased.

[romainfrancois]

Obsolete now, as #3635 was merged.