More robust DMatrix creation from a sparse matrix (#1606)

* [CORE] DMatrix from sparse w/ explicit #col #row; safer arg types

* [python-package] c-api change for _init_from_csr _init_from_csc

* fix spaces

* [R-package] adopt the new XGDMatrixCreateFromCSCEx interface

* [CORE] redirect old sparse creators to new ones

R-package/tests/testthat/test_dmatrix.R

@@ -1,4 +1,5 @@
 require(xgboost)
+require(Matrix)
 
 context("testing xgb.DMatrix functionality")
 
@@ -65,3 +66,13 @@ test_that("xgb.DMatrix: colnames", {
   expect_silent(colnames(dtest) <- NULL)
   expect_null(colnames(dtest))
 })
+
+test_that("xgb.DMatrix: nrow is correct for a very sparse matrix", {
+  set.seed(123)
+  nr <- 1000
+  x <- rsparsematrix(nr, 100, density=0.0005)
+  # we want it very sparse, so that last rows are empty
+  expect_lt(max(x@i), nr)
+  dtest <- xgb.DMatrix(x)
+  expect_equal(dim(dtest), dim(x))
+})