The result of an Rserve get command may be a SEXP expression - a recursively encoded list of names and values. Such expressions should be converted to a more convenient form in J.
There are three forms in which the SEXP may be returned:
- Rcmdrexp returns the original SEXP
- Rcmdr returns a map (the usual format)
- Rcmdrtree returns a tree.
1. Rcmdrexp returns the result in the original SEXP format. This is useful for debugging, but inconvenient for practical use:
Rcmdrexp 'x=factor(c("one","two","three","four"))'
┌────────────────────────────────────────────┬───────┐
│┌────────────────────┬───────┬──────┬──────┐│2 4 3 1│
││┌────┬───┬─────┬───┐│`levels│factor│`class││ │
│││four│one│three│two││ │ │ ││ │
││└────┴───┴─────┴───┘│ │ │ ││ │
│└────────────────────┴───────┴──────┴──────┘│ │
└────────────────────────────────────────────┴───────┘ 2. Rcmdr returns the result as a map, i.e. a 2-column matrix of names and values:
Rcmdr 'x' ┌──────┬────────────────────┐ │class │factor │ ├──────┼────────────────────┤ │data │2 4 3 1 │ ├──────┼────────────────────┤ │levels│┌────┬───┬─────┬───┐│ │ ││four│one│three│two││ │ │└────┴───┴─────┴───┘│ └──────┴────────────────────┘
The name 'data' is used where the name is not given.
Where the SEXP is nested, the name paths are separated by dots:
Rcmd 'data(OrchardSprays)' Rcmdr 'OrchardSprays' ┌────────────────┬──────────────────────────────────────────────────────────────... │class │data.frame ... ├────────────────┼──────────────────────────────────────────────────────────────... │colpos │1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 ... ├────────────────┼──────────────────────────────────────────────────────────────... │decrease │57 95 8 69 92 90 15 2 84 6 127 36 51 2 69 71 87 72 5 39 22 16 ... ├────────────────┼──────────────────────────────────────────────────────────────... │rownames │1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24... ├────────────────┼──────────────────────────────────────────────────────────────... │rowpos │1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 ... ├────────────────┼──────────────────────────────────────────────────────────────... │treatment │4 5 2 8 7 6 3 1 3 2 8 4 5 1 6 7 6 8 1 5 4 3 7 2 8 1 5 3 6 7 2 ... ├────────────────┼──────────────────────────────────────────────────────────────... │treatment.class │factor ... ├────────────────┼──────────────────────────────────────────────────────────────... │treatment.levels│┌─┬─┬─┬─┬─┬─┬─┬─┐ ... │ ││A│B│C│D│E│F│G│H│ ... │ │└─┴─┴─┴─┴─┴─┴─┴─┘ ... └────────────────┴──────────────────────────────────────────────────────────────...
In nearly all cases, the map will be the most useful result. However, there are a couple of problems with this format:
- The names in the first column do not distinguish attribute names from data names.
- Some R names have a dot separator, e.g. 'row.names', even though they appear in Rserve as single identifiers - i.e. there is no item 'row' with a 'names' attribute. This is confusing for the map format, and as a result the '.' separator is removed from such names.
In practice, these issues should not be a problem for a J program that is driving an R server, since the program can avoid using names that clash wih R internal names. In the rare case that the J result needs to be completely unambiguous, the tree format can be used.
3. Rcmdrtree returns the result as a tree. In this case, the J result exactly matches the original SEXP.
In the tree, any R identifiers are given as symbols. Any nesting is handled by boxing the data. For example:
Rcmdrtree 'OrchardSprays' ┌──────────┬────────────────────────────────────────────────────────────────────... │`class │data.frame ... ├──────────┼────────────────────────────────────────────────────────────────────... │`row.names│1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26... ├──────────┼────────────────────────────────────────────────────────────────────... │colpos │1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 4 4 4 4 4 4 4 4 5 5 ... ├──────────┼────────────────────────────────────────────────────────────────────... │decrease │57 95 8 69 92 90 15 2 84 6 127 36 51 2 69 71 87 72 5 39 22 16 72 4 1... ├──────────┼────────────────────────────────────────────────────────────────────... │rowpos │1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 ... ├──────────┼────────────────────────────────────────────────────────────────────... │treatment │┌───────┬───────────────────────────────────────────────────────────... │ ││`data │4 5 2 8 7 6 3 1 3 2 8 4 5 1 6 7 6 8 1 5 4 3 7 2 8 1 5 3 6 7... │ │├───────┼───────────────────────────────────────────────────────────... │ ││`levels│┌─┬─┬─┬─┬─┬─┬─┬─┐ ... │ ││ ││A│B│C│D│E│F│G│H│ ... │ ││ │└─┴─┴─┴─┴─┴─┴─┴─┘ ... │ │├───────┼───────────────────────────────────────────────────────────... │ ││`class │factor ... │ │└───────┴───────────────────────────────────────────────────────────... └──────────┴────────────────────────────────────────────────────────────────────...