Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[BLOCKING] Handle empty rows in data iterators correctly (#5929)

Browse Source 
* [jvm-packages] Handle empty rows in data iterators correctly

* Fix clang-tidy error

* last empty row

* Add comments [skip ci]

Co-authored-by: Nan Zhu <nanzhu@uber.com>
This commit is contained in:
Philip Hyunsu Cho
2020-07-25 13:46:19 -07:00
committed by GitHub
parent a4de2f68e4
commit 487ab0ce73
5 changed files with 79 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
tests/cpp/data/test_simple_dmatrix.cc
View File

@@ -185,16 +185,22 @@ TEST(SimpleDMatrix, FromCSC) {
TEST(SimpleDMatrix, FromFile) {
std::string filename = "test.libsvm";
CreateBigTestData(filename, 3 * 5);
// Add an empty row at the end of the matrix
{
std::ofstream fo(filename, std::ios::app | std::ios::out);
fo << "0\n";
}
constexpr size_t kExpectedNumRow = 6;
std::unique_ptr<dmlc::Parser<uint32_t>> parser(
dmlc::Parser<uint32_t>::Create(filename.c_str(), 0, 1, "auto"));
auto verify_batch = [](SparsePage const &batch) {
EXPECT_EQ(batch.Size(), 5);
auto verify_batch = [kExpectedNumRow](SparsePage const &batch) {
EXPECT_EQ(batch.Size(), kExpectedNumRow);
EXPECT_EQ(batch.offset.HostVector(),
std::vector<bst_row_t>({0, 3, 6, 9, 12, 15}));
std::vector<bst_row_t>({0, 3, 6, 9, 12, 15, 15}));
EXPECT_EQ(batch.base_rowid, 0);
for (auto i = 0ull; i < batch.Size(); i++) {
for (auto i = 0ull; i < batch.Size() - 1; i++) {
if (i % 2 == 0) {
EXPECT_EQ(batch[i][0].index, 0);
EXPECT_EQ(batch[i][1].index, 1);