Skip to content

Commit d5528f9

Browse files
WillAydproost
authored andcommitted
use stdlib for whitespace check (pandas-dev#30061)
1 parent 77698c8 commit d5528f9

File tree

1 file changed

+4
-6
lines changed

1 file changed

+4
-6
lines changed

pandas/_libs/src/parser/tokenizer.c

+4-6
Original file line numberDiff line numberDiff line change
@@ -647,8 +647,6 @@ static int parser_buffer_bytes(parser_t *self, size_t nbytes) {
647647

648648
#define END_LINE() END_LINE_STATE(START_RECORD)
649649

650-
#define IS_WHITESPACE(c) ((c == ' ' || c == '\t'))
651-
652650
#define IS_TERMINATOR(c) \
653651
(c == line_terminator)
654652

@@ -667,7 +665,7 @@ static int parser_buffer_bytes(parser_t *self, size_t nbytes) {
667665
// applied when in a field
668666
#define IS_DELIMITER(c) \
669667
((!self->delim_whitespace && c == self->delimiter) || \
670-
(self->delim_whitespace && IS_WHITESPACE(c)))
668+
(self->delim_whitespace && isblank(c)))
671669

672670
#define _TOKEN_CLEANUP() \
673671
self->stream_len = slen; \
@@ -818,7 +816,7 @@ int tokenize_bytes(parser_t *self,
818816
self->state = EAT_CRNL_NOP;
819817
break;
820818
} else if (!self->delim_whitespace) {
821-
if (IS_WHITESPACE(c) && c != self->delimiter) {
819+
if (isblank(c) && c != self->delimiter) {
822820
} else { // backtrack
823821
// use i + 1 because buf has been incremented but not i
824822
do {
@@ -848,7 +846,7 @@ int tokenize_bytes(parser_t *self,
848846
} else if (IS_COMMENT_CHAR(c)) {
849847
self->state = EAT_COMMENT;
850848
break;
851-
} else if (!IS_WHITESPACE(c)) {
849+
} else if (!isblank(c)) {
852850
self->state = START_FIELD;
853851
// fall through to subsequent state
854852
} else {
@@ -892,7 +890,7 @@ int tokenize_bytes(parser_t *self,
892890
} else if (IS_COMMENT_CHAR(c)) {
893891
self->state = EAT_LINE_COMMENT;
894892
break;
895-
} else if (IS_WHITESPACE(c)) {
893+
} else if (isblank(c)) {
896894
if (self->delim_whitespace) {
897895
if (self->skip_empty_lines) {
898896
self->state = WHITESPACE_LINE;

0 commit comments

Comments
 (0)