jinchengchenghh commented on code in PR #43482:
URL: https://github.com/apache/arrow/pull/43482#discussion_r1701065457
##########
java/dataset/src/main/cpp/jni_wrapper.cc:
##########
@@ -368,28 +368,105 @@ std::shared_ptr<arrow::Buffer>
LoadArrowBufferFromByteBuffer(JNIEnv* env, jobjec
inline bool ParseBool(const std::string& value) { return value == "true" ?
true : false; }
+inline bool ParseChar(const std::string& value) {
+ if (value.size() != 1) {
+ JniThrow("Csv convert option " + value + " should be a char");
+ }
+ return value.at(0);
+}
+
/// \brief Construct FragmentScanOptions from config map
#ifdef ARROW_CSV
+
+bool SetCsvConvertOptions(arrow::csv::ConvertOptions& options, const
std::string& key,
+ const std::string& value) {
+ if (key == "column_types") {
+ int64_t schema_address = std::stol(value);
+ ArrowSchema* c_schema = reinterpret_cast<ArrowSchema*>(schema_address);
+ auto schema = JniGetOrThrow(arrow::ImportSchema(c_schema));
+ auto& column_types = options.column_types;
+ for (auto field : schema->fields()) {
+ column_types[field->name()] = field->type();
+ }
+ } else if (key == "strings_can_be_null") {
+ options.strings_can_be_null = ParseBool(value);
+ } else if (key == "check_utf8") {
+ options.check_utf8 = ParseBool(value);
+ } else if (key == "null_values") {
+ options.null_values = {value};
+ } else if (key == "true_values") {
+ options.true_values = {value};
+ } else if (key == "false_values") {
+ options.false_values = {value};
+ } else if (key == "quoted_strings_can_be_null") {
+ options.quoted_strings_can_be_null = ParseBool(value);
+ } else if (key == "auto_dict_encode") {
+ options.auto_dict_encode = ParseBool(value);
+ } else if (key == "auto_dict_max_cardinality") {
+ options.auto_dict_max_cardinality = std::stoi(value);
+ } else if (key == "decimal_point") {
+ options.decimal_point = ParseChar(value);
+ } else if (key == "include_missing_columns") {
+ options.include_missing_columns = ParseBool(value);
+ } else {
+ return false;
+ }
+ return true;
+}
+
+bool SetCsvParseOptions(arrow::csv::ParseOptions& options, const std::string&
key,
+ const std::string& value) {
+ if (key == "delimiter") {
+ options.delimiter = ParseChar(value);
+ } else if (key == "quoting") {
+ options.quoting = ParseBool(value);
+ } else if (key == "quote_char") {
+ options.quote_char = ParseChar(value);
+ } else if (key == "double_quote") {
+ options.double_quote = ParseBool(value);
+ } else if (key == "escaping") {
+ options.escaping = ParseBool(value);
+ } else if (key == "escape_char") {
+ options.escape_char = ParseChar(value);
+ } else if (key == "newlines_in_values") {
+ options.newlines_in_values = ParseBool(value);
+ } else if (key == "ignore_empty_lines") {
+ options.ignore_empty_lines = ParseBool(value);
+ } else {
+ return false;
+ }
+ return true;
+}
+
+bool SetCsvReadOptions(arrow::csv::ReadOptions& options, const std::string&
key,
+ const std::string& value) {
+ if (key == "use_threads") {
+ options.use_threads = ParseBool(value);
+ } else if (key == "block_size") {
+ options.block_size = std::stoi(value);
+ } else if (key == "skip_rows") {
+ options.skip_rows = std::stoi(value);
+ } else if (key == "skip_rows_after_names") {
+ options.skip_rows_after_names = std::stoi(value);
+ } else if (key == "autogenerate_column_names") {
+ options.autogenerate_column_names = ParseBool(value);
+ } else {
+ return false;
+ }
+ return true;
+}
+
arrow::Result<std::shared_ptr<arrow::dataset::FragmentScanOptions>>
ToCsvFragmentScanOptions(const std::unordered_map<std::string, std::string>&
configs) {
std::shared_ptr<arrow::dataset::CsvFragmentScanOptions> options =
std::make_shared<arrow::dataset::CsvFragmentScanOptions>();
- for (auto const& [key, value] : configs) {
- if (key == "delimiter") {
- options->parse_options.delimiter = value.data()[0];
- } else if (key == "quoting") {
- options->parse_options.quoting = ParseBool(value);
- } else if (key == "column_types") {
- int64_t schema_address = std::stol(value);
- ArrowSchema* c_schema = reinterpret_cast<ArrowSchema*>(schema_address);
- ARROW_ASSIGN_OR_RAISE(auto schema, arrow::ImportSchema(c_schema));
- auto& column_types = options->convert_options.column_types;
- for (auto field : schema->fields()) {
- column_types[field->name()] = field->type();
- }
- } else if (key == "strings_can_be_null") {
- options->convert_options.strings_can_be_null = ParseBool(value);
- } else {
+ for (const auto& it : configs) {
Review Comment:
No, I meet exception, so I change to current.
```
/arrow/java/dataset/src/main/cpp/jni_wrapper.cc:464:69: error: binding
reference of type ‘std::string&’ {aka ‘std::basic_string<char>&’} to
‘std::tuple_element<1, const std::pair<const std::basic_string<char>,
std::basic_string<char> > >::type’ {aka ‘const std::basic_string<char>’}
discards qualifiers
464 | bool setValid = setCsvParseOptions(options->parse_options, key,
value) &&
|
^~~~~
/arrow/java/dataset/src/main/cpp/jni_wrapper.cc:418:38: note: initializing
argument 3 of ‘bool setCsvParseOptions(arrow::csv::ParseOptions&, const
string&, std::string&)’
418 | std::string& value) {
|
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
