[GitHub] [incubator-doris] morningman commented on a change in pull request #5462: [Load] Support multi bytes LineDelimiter and ColumnSeparator

[GitBox <gi...@apache.org>]

morningman commented on a change in pull request #5462:
URL: https://github.com/apache/incubator-doris/pull/5462#discussion_r587521289



##########
File path: be/src/exec/broker_scanner.cpp
##########
@@ -292,16 +304,24 @@ void BrokerScanner::close() {
 }
 
 void BrokerScanner::split_line(const Slice& line, std::vector<Slice>* values) {
-    // line-begin char and line-end char are considered to be 'delimiter'
     const char* value = line.data;
-    const char* ptr = line.data;
-    for (size_t i = 0; i < line.size; ++i, ++ptr) {
-        if (*ptr == _value_separator) {
-            values->emplace_back(value, ptr - value);
-            value = ptr + 1;
+    size_t i = 0;
+    // TODO improve the performance
+    while (i < line.size) {
+        if (i + _value_separator.size() <= line.size) {

Review comment:
       Does `_value_separator.size()` equals to `_value_separator_length`?

##########
File path: be/src/exec/broker_scanner.cpp
##########
@@ -49,16 +49,28 @@ BrokerScanner::BrokerScanner(RuntimeState* state, RuntimeProfile* profile,
         : BaseScanner(state, profile, params, pre_filter_ctxs, counter),
           _ranges(ranges),
           _broker_addresses(broker_addresses),
-          // _splittable(params.splittable),
-          _value_separator(static_cast<char>(params.column_separator)),
-          _line_delimiter(static_cast<char>(params.line_delimiter)),
           _cur_file_reader(nullptr),
           _cur_line_reader(nullptr),
           _cur_decompressor(nullptr),
           _next_range(0),
           _cur_line_reader_eof(false),
           _scanner_eof(false),
-          _skip_next_line(false) {}
+          _skip_next_line(false) {
+    if (params.__isset.column_separator_length && params.column_separator_length > 1) {
+        _value_separator = params.column_separator_str;
+        _value_separator_length = params.column_separator_length;
+    } else {
+        _value_separator.push_back(static_cast<char>(params.column_separator));
+        _value_separator_length = 1;
+    }
+    if (params.__isset.line_delimiter_length && params.line_delimiter_length > 1) {
+        _line_delimiter = params.line_delimiter_str;
+        _line_delimiter_length = params.line_delimiter_length;
+    } else {
+        _line_delimiter.push_back(static_cast<char>(params.line_delimiter));
+        _line_delimiter_length = 1;

Review comment:
       `_line_delimiter_length` seams useless.




----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@doris.apache.org
For additional commands, e-mail: commits-help@doris.apache.org