Skip to content

Commit

Permalink
apacheGH-38090: [C++][Emscripten] acero/tpch_node: Suppress shorten-6…
Browse files Browse the repository at this point in the history 
…4-to-32 warnings

We need explicit cast to use `int64_t` for `size_t` on Emscripten.

Explicit casts.
  • Loading branch information
@kou
kou committed Oct 6, 2023
1 parent 3697bcd commit 8285399
Showing 1 changed file with 49 additions and 41 deletions.
90 changes: 49 additions & 41 deletions cpp/src/arrow/acero/tpch_node.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -307,7 +307,7 @@ Status TpchPseudotext::EnsureInitialized(random::pcg32_fast& rng) {
generated_offset_ += memcpy_size;
last_one = generated_offset_ == kTextBytes;
}
std::memcpy(out + offset, temp_buff, memcpy_size);
std::memcpy(out + offset, temp_buff, static_cast<size_t>(memcpy_size));
if (last_one) done_.store(true);
}
return Status::OK();
Expand Down Expand Up @@ -646,7 +646,8 @@ class PartAndPartSupplierGenerator {
random::pcg64_fast seed_rng(seed);
for (ThreadLocalData& tld : thread_local_data_) {
constexpr int kMaxNumDistinctStrings = 5;
tld.string_indices.resize(kMaxNumDistinctStrings * batch_size_);
tld.string_indices.resize(
static_cast<size_t>(kMaxNumDistinctStrings * batch_size_));
tld.rng.seed(kSeedDist(seed_rng));
}
part_rows_to_generate_ = static_cast<int64_t>(scale_factor_ * 200000);
Expand Down Expand Up @@ -880,9 +881,9 @@ class PartAndPartSupplierGenerator {
AllocateBuffer((tld.part_to_generate + 1) * sizeof(int32_t)));
int32_t* offsets = reinterpret_cast<int32_t*>(offset_buff->mutable_data());
offsets[0] = 0;
for (int64_t irow = 0; irow < tld.part_to_generate; irow++) {
for (size_t irow = 0; irow < static_cast<size_t>(tld.part_to_generate); irow++) {
size_t string_length = 0;
for (int ipart = 0; ipart < 5; ipart++) {
for (size_t ipart = 0; ipart < 5; ipart++) {
uint8_t name_part_index = static_cast<uint8_t>(dist(tld.rng));
tld.string_indices[irow * 5 + ipart] = name_part_index;
string_length += std::strlen(NameParts[name_part_index]);
Expand All @@ -894,9 +895,9 @@ class PartAndPartSupplierGenerator {
ARROW_ASSIGN_OR_RAISE(std::unique_ptr<Buffer> string_buffer,
AllocateBuffer(offsets[tld.part_to_generate] + 1));
char* strings = reinterpret_cast<char*>(string_buffer->mutable_data());
for (int64_t irow = 0; irow < tld.part_to_generate; irow++) {
for (size_t irow = 0; irow < static_cast<size_t>(tld.part_to_generate); irow++) {
char* row = strings + offsets[irow];
for (int ipart = 0; ipart < 5; ipart++) {
for (size_t ipart = 0; ipart < 5; ipart++) {
uint8_t name_part_index = tld.string_indices[irow * 5 + ipart];
const char* part = NameParts[name_part_index];
size_t length = std::strlen(part);
Expand Down Expand Up @@ -976,9 +977,9 @@ class PartAndPartSupplierGenerator {
AllocateBuffer((tld.part_to_generate + 1) * sizeof(int32_t)));
int32_t* offsets = reinterpret_cast<int32_t*>(offset_buff->mutable_data());
offsets[0] = 0;
for (int64_t irow = 0; irow < tld.part_to_generate; irow++) {
for (size_t irow = 0; irow < static_cast<size_t>(tld.part_to_generate); irow++) {
size_t string_length = 0;
for (int ipart = 0; ipart < 3; ipart++) {
for (size_t ipart = 0; ipart < 3; ipart++) {
uint8_t name_part_index = static_cast<uint8_t>(dists[ipart](tld.rng));
tld.string_indices[irow * 3 + ipart] = name_part_index;
string_length += std::strlen(types[ipart][name_part_index]);
Expand All @@ -988,9 +989,9 @@ class PartAndPartSupplierGenerator {
ARROW_ASSIGN_OR_RAISE(std::unique_ptr<Buffer> string_buffer,
AllocateBuffer(offsets[tld.part_to_generate]));
char* strings = reinterpret_cast<char*>(string_buffer->mutable_data());
for (int64_t irow = 0; irow < tld.part_to_generate; irow++) {
for (size_t irow = 0; irow < static_cast<size_t>(tld.part_to_generate); irow++) {
char* row = strings + offsets[irow];
for (int ipart = 0; ipart < 3; ipart++) {
for (size_t ipart = 0; ipart < 3; ipart++) {
uint8_t name_part_index = tld.string_indices[irow * 3 + ipart];
const char* part = types[ipart][name_part_index];
size_t length = std::strlen(part);
Expand Down Expand Up @@ -1069,7 +1070,8 @@ class PartAndPartSupplierGenerator {
if (tld.part[PART::P_COMMENT].kind() == Datum::NONE) {
ARROW_ASSIGN_OR_RAISE(
tld.part[PART::P_COMMENT],
g_text.GenerateComments(tld.part_to_generate, 5, 22, tld.rng));
g_text.GenerateComments(static_cast<size_t>(tld.part_to_generate), 5, 22,
tld.rng));
}
return Status::OK();
}
Expand All @@ -1085,7 +1087,7 @@ class PartAndPartSupplierGenerator {
ThreadLocalData& tld = thread_local_data_[thread_index];
tld.generated_partsupp.reset();
int64_t num_batches = PartsuppBatchesToGenerate(thread_index);
tld.partsupp.resize(num_batches);
tld.partsupp.resize(static_cast<size_t>(num_batches));
for (std::vector<Datum>& batch : tld.partsupp) {
batch.resize(PARTSUPP::kNumCols);
std::fill(batch.begin(), batch.end(), Datum());
Expand Down Expand Up @@ -1145,8 +1147,8 @@ class PartAndPartSupplierGenerator {
ipart++;
}
}
RETURN_NOT_OK(
SetPartSuppColumnSize(thread_index, ibatch, PARTSUPP::PS_PARTKEY, next_run));
RETURN_NOT_OK(SetPartSuppColumnSize(thread_index, ibatch, PARTSUPP::PS_PARTKEY,
static_cast<size_t>(next_run)));
irow += next_run;
}
}
Expand Down Expand Up @@ -1190,8 +1192,8 @@ class PartAndPartSupplierGenerator {
ipart++;
}
}
RETURN_NOT_OK(
SetPartSuppColumnSize(thread_index, ibatch, PARTSUPP::PS_SUPPKEY, next_run));
RETURN_NOT_OK(SetPartSuppColumnSize(thread_index, ibatch, PARTSUPP::PS_SUPPKEY,
static_cast<size_t>(next_run)));
irow += next_run;
}
}
Expand All @@ -1203,17 +1205,19 @@ class PartAndPartSupplierGenerator {
if (!tld.generated_partsupp[PARTSUPP::PS_AVAILQTY]) {
tld.generated_partsupp[PARTSUPP::PS_AVAILQTY] = true;
std::uniform_int_distribution<int32_t> dist(1, 9999);
int64_t ps_to_generate = kPartSuppRowsPerPart * tld.part_to_generate;
int64_t ibatch = 0;
for (int64_t irow = 0; irow < ps_to_generate; ibatch++) {
auto ps_to_generate =
static_cast<size_t>(kPartSuppRowsPerPart * tld.part_to_generate);
size_t ibatch = 0;
for (size_t irow = 0; irow < ps_to_generate; ibatch++) {
RETURN_NOT_OK(AllocatePartSuppBatch(thread_index, ibatch, PARTSUPP::PS_AVAILQTY));
int32_t* ps_availqty =
reinterpret_cast<int32_t*>(tld.partsupp[ibatch][PARTSUPP::PS_AVAILQTY]
.array()
->buffers[1]
->mutable_data());
int64_t next_run = std::min(batch_size_, ps_to_generate - irow);
for (int64_t irun = 0; irun < next_run; irun++) ps_availqty[irun] = dist(tld.rng);
size_t next_run =
std::min(static_cast<size_t>(batch_size_), ps_to_generate - irow);
for (size_t irun = 0; irun < next_run; irun++) ps_availqty[irun] = dist(tld.rng);

RETURN_NOT_OK(
SetPartSuppColumnSize(thread_index, ibatch, PARTSUPP::PS_AVAILQTY, next_run));
Expand All @@ -1229,7 +1233,7 @@ class PartAndPartSupplierGenerator {
tld.generated_partsupp[PARTSUPP::PS_SUPPLYCOST] = true;
std::uniform_int_distribution<int64_t> dist(100, 100000);
int64_t ps_to_generate = kPartSuppRowsPerPart * tld.part_to_generate;
int64_t ibatch = 0;
size_t ibatch = 0;
for (int64_t irow = 0; irow < ps_to_generate; ibatch++) {
RETURN_NOT_OK(
AllocatePartSuppBatch(thread_index, ibatch, PARTSUPP::PS_SUPPLYCOST));
Expand All @@ -1238,8 +1242,8 @@ class PartAndPartSupplierGenerator {
.array()
->buffers[1]
->mutable_data());
int64_t next_run = std::min(batch_size_, ps_to_generate - irow);
for (int64_t irun = 0; irun < next_run; irun++)
auto next_run = static_cast<size_t>(std::min(batch_size_, ps_to_generate - irow));
for (size_t irun = 0; irun < next_run; irun++)
ps_supplycost[irun] = {dist(tld.rng)};

RETURN_NOT_OK(SetPartSuppColumnSize(thread_index, ibatch, PARTSUPP::PS_SUPPLYCOST,
Expand All @@ -1257,7 +1261,7 @@ class PartAndPartSupplierGenerator {
int64_t irow = 0;
int64_t ps_to_generate = kPartSuppRowsPerPart * tld.part_to_generate;
for (size_t ibatch = 0; ibatch < tld.partsupp.size(); ibatch++) {
int64_t num_rows = std::min(batch_size_, ps_to_generate - irow);
auto num_rows = static_cast<size_t>(std::min(batch_size_, ps_to_generate - irow));
ARROW_ASSIGN_OR_RAISE(tld.partsupp[ibatch][PARTSUPP::PS_COMMENT],
g_text.GenerateComments(num_rows, 49, 198, tld.rng));
irow += num_rows;
Expand Down Expand Up @@ -1306,7 +1310,7 @@ class OrdersAndLineItemGenerator {
thread_local_data_.resize(num_threads);
random::pcg64_fast seed_rng(seed);
for (ThreadLocalData& tld : thread_local_data_) {
tld.items_per_order.resize(batch_size_);
tld.items_per_order.resize(static_cast<size_t>(batch_size_));
tld.rng.seed(kSeedDist(seed_rng));
}
orders_rows_to_generate_ = static_cast<int64_t>(scale_factor_ * 150000 * 10);
Expand Down Expand Up @@ -1348,7 +1352,7 @@ class OrdersAndLineItemGenerator {
orders_batches_generated_.fetch_add(1);
tld.first_batch_offset = 0;
RETURN_NOT_OK(GenerateRowCounts(thread_index));
lineitem_batches_generated_.fetch_add(static_cast<int64_t>(tld.lineitem.size()));
lineitem_batches_generated_.fetch_add(tld.lineitem.size());
ARROW_DCHECK(orders_rows_generated_ <= orders_rows_to_generate_);
}
}
Expand Down Expand Up @@ -1416,8 +1420,7 @@ class OrdersAndLineItemGenerator {
orders_rows_generated_ += tld.orders_to_generate;
orders_batches_generated_.fetch_add(1);
RETURN_NOT_OK(GenerateRowCounts(thread_index));
lineitem_batches_generated_.fetch_add(
static_cast<int64_t>(tld.lineitem.size() - from_queue));
lineitem_batches_generated_.fetch_add(tld.lineitem.size() - from_queue);
ARROW_DCHECK(orders_rows_generated_ <= orders_rows_to_generate_);
}
tld.orders.resize(ORDERS::kNumCols);
Expand Down Expand Up @@ -1611,7 +1614,7 @@ class OrdersAndLineItemGenerator {
char* o_orderstatus = reinterpret_cast<char*>(
tld.orders[ORDERS::O_ORDERSTATUS].array()->buffers[1]->mutable_data());

size_t batch_offset = tld.first_batch_offset;
auto batch_offset = static_cast<size_t>(tld.first_batch_offset);
size_t ibatch = 0;
size_t iorder = 0;
int32_t iline = 0;
Expand Down Expand Up @@ -1656,7 +1659,7 @@ class OrdersAndLineItemGenerator {
RETURN_NOT_OK(L_DISCOUNT(thread_index));
RETURN_NOT_OK(AllocateOrdersBatch(thread_index, ORDERS::O_TOTALPRICE));

size_t batch_offset = tld.first_batch_offset;
auto batch_offset = static_cast<size_t>(tld.first_batch_offset);
size_t ibatch = 0;
size_t iorder = 0;
int32_t iline = 0;
Expand Down Expand Up @@ -1752,7 +1755,8 @@ class OrdersAndLineItemGenerator {
RETURN_NOT_OK(AllocateOrdersBatch(thread_index, ORDERS::O_SHIPPRIORITY));
int32_t* o_shippriority = reinterpret_cast<int32_t*>(
tld.orders[ORDERS::O_SHIPPRIORITY].array()->buffers[1]->mutable_data());
std::memset(o_shippriority, 0, tld.orders_to_generate * sizeof(int32_t));
std::memset(o_shippriority, 0,
static_cast<size_t>(tld.orders_to_generate) * sizeof(int32_t));
}
return Status::OK();
}
Expand All @@ -1762,7 +1766,8 @@ class OrdersAndLineItemGenerator {
if (tld.orders[ORDERS::O_COMMENT].kind() == Datum::NONE) {
ARROW_ASSIGN_OR_RAISE(
tld.orders[ORDERS::O_COMMENT],
g_text.GenerateComments(tld.orders_to_generate, 19, 78, tld.rng));
g_text.GenerateComments(static_cast<size_t>(tld.orders_to_generate), 19, 78,
tld.rng));
}
return Status::OK();
}
Expand All @@ -1777,9 +1782,9 @@ class OrdersAndLineItemGenerator {
tld.items_per_order.push_back(length);
tld.lineitem_to_generate += length;
}
int64_t num_batches =
auto num_batches = static_cast<size_t>(
(tld.first_batch_offset + tld.lineitem_to_generate + batch_size_ - 1) /
batch_size_;
batch_size_);
tld.lineitem.resize(num_batches);
for (std::vector<Datum>& batch : tld.lineitem) {
batch.resize(LINEITEM::kNumCols);
Expand All @@ -1800,7 +1805,7 @@ class OrdersAndLineItemGenerator {
tld.lineitem[ibatch][column] = std::move(ad);
out_batch_offset = 0;
}
if (ibatch == 0) out_batch_offset = tld.first_batch_offset;
if (ibatch == 0) out_batch_offset = static_cast<size_t>(tld.first_batch_offset);

return Status::OK();
}
Expand Down Expand Up @@ -2357,11 +2362,12 @@ class OrdersAndLineItemGenerator {
// bump the length
size_t batch_offset = 0;
if (tld.lineitem[ibatch][LINEITEM::L_COMMENT].kind() == Datum::NONE) {
ARROW_ASSIGN_OR_RAISE(tld.lineitem[ibatch][LINEITEM::L_COMMENT],
g_text.GenerateComments(batch_size_, 10, 43, tld.rng));
ARROW_ASSIGN_OR_RAISE(
tld.lineitem[ibatch][LINEITEM::L_COMMENT],
g_text.GenerateComments(static_cast<size_t>(batch_size_), 10, 43, tld.rng));
batch_offset = 0;
}
if (irow == 0) batch_offset = tld.first_batch_offset;
if (irow == 0) batch_offset = static_cast<size_t>(tld.first_batch_offset);

int64_t remaining_in_batch = static_cast<int64_t>(batch_size_ - batch_offset);
int64_t next_run = std::min(tld.lineitem_to_generate - irow, remaining_in_batch);
Expand Down Expand Up @@ -2627,7 +2633,8 @@ class SupplierGenerator : public TpchTableGenerator {
ThreadLocalData& tld = thread_local_data_[thread_index];
if (tld.batch[SUPPLIER::S_COMMENT].kind() == Datum::NONE) {
ARROW_ASSIGN_OR_RAISE(tld.batch[SUPPLIER::S_COMMENT],
g_text.GenerateComments(tld.to_generate, 25, 100, tld.rng));
g_text.GenerateComments(static_cast<size_t>(tld.to_generate),
25, 100, tld.rng));
ModifyComments(thread_index, "Recommends", good_rows_);
ModifyComments(thread_index, "Complaints", bad_rows_);
}
Expand Down Expand Up @@ -3039,7 +3046,8 @@ class CustomerGenerator : public TpchTableGenerator {
ThreadLocalData& tld = thread_local_data_[thread_index];
if (tld.batch[CUSTOMER::C_COMMENT].kind() == Datum::NONE) {
ARROW_ASSIGN_OR_RAISE(tld.batch[CUSTOMER::C_COMMENT],
g_text.GenerateComments(tld.to_generate, 29, 116, tld.rng));
g_text.GenerateComments(static_cast<size_t>(tld.to_generate),
29, 116, tld.rng));
}
return Status::OK();
}
Expand Down

0 comments on commit 8285399

Please sign in to comment.