Skip to content

GH-54534: [C++] Fix truncation of sub-minute UTC offsets when formatting and parsing zoned timestamps #54684

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
ShivanshhhG wants to merge 1 commit into
base: main
Choose a base branch
from
+99 −19
Open

GH-54534: [C++] Fix truncation of sub-minute UTC offsets when formatting and parsing zoned timestamps #54684

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 10 additions & 0 deletions cpp/src/arrow/compute/kernels/scalar_cast_test.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2792,6 +2792,16 @@ TEST_F(CastTimezone, TimestampWithZoneToString) {
ArrayFromJSON(
string_type,
R"(["1968-11-30 13:30:44.123456789-0700", "2016-02-29 10:42:23.456789246-0700"])"));

CheckCast(ArrayFromJSON(timestamp(TimeUnit::MICRO, "Africa/Monrovia"),
"[-302443200000000]"),
ArrayFromJSON(string_type,
R"(["1960-06-01 11:15:30.000000-004430"])"));

CheckCast(ArrayFromJSON(timestamp(TimeUnit::MICRO, "Europe/Paris"),
"[-2195899200000000]"),
ArrayFromJSON(string_type,
R"(["1900-06-01 12:09:21.000000+000921"])"));
}
}

Expand Down
11 changes: 7 additions & 4 deletions cpp/src/arrow/compute/kernels/scalar_temporal_unary.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1262,16 +1262,19 @@ struct Strftime {
// Convert string representations of timestamps in arbitrary format to timestamps

const std::string GetZone(const std::string& format) {
// Check for use of %z or %Z
// Check for use of %z or %Ez

@dmitry-chirkov-dremio dmitry-chirkov-dremio Oct 22, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

%Ez is now recognized here and in StrptimeTimestampParser, but Strftime::Make still checks only %z and %Z when the input has no timezone. This means %Ez is treated as UTC while %z is rejected. Please use the same format check in both paths and add tests for %Ez with and without a timezone.

size_t cur = 0;
size_t count = 0;
std::string zone = "";
while (cur < format.size() - 1) {
if (format[cur] == '%') {
count++;
if (format[cur + 1] == 'z' && count % 2 == 1) {
zone = "UTC";
break;
if (count % 2 == 1) {
if (format[cur + 1] == 'z' ||
(format[cur + 1] == 'E' && cur + 2 < format.size() && format[cur + 2] == 'z')) {
zone = "UTC";
break;
}
}
} else {
count = 0;
Expand Down
7 changes: 5 additions & 2 deletions cpp/src/arrow/util/value_parsing.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -78,11 +78,14 @@ class StrptimeTimestampParser : public TimestampParser {
public:
explicit StrptimeTimestampParser(std::string format)
: format_(std::move(format)), have_zone_offset_(false) {
// Check for use of %z
// Check for use of %z or %Ez
size_t cur = 0;
while (cur < format_.size()) {
if (format_[cur] == '%') {
if (cur + 1 < format_.size() && format_[cur + 1] == 'z') {
if (cur + 1 < format_.size() &&
(format_[cur + 1] == 'z' ||
(format_[cur + 1] == 'E' && cur + 2 < format_.size() &&
format_[cur + 2] == 'z'))) {
have_zone_offset_ = true;
break;
}
Expand Down
67 changes: 57 additions & 10 deletions cpp/src/arrow/util/value_parsing.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -553,6 +553,35 @@ static inline bool ParseHH_MM_SS(const char* s, Duration* out) {
return true;
}

template <typename Duration>
static inline bool ParseHHMMSS(const char* s, Duration* out) {
uint8_t hours = 0;
uint8_t minutes = 0;
uint8_t seconds = 0;
if (ARROW_PREDICT_FALSE(!ParseUnsigned(s + 0, 2, &hours))) {
return false;
}
if (ARROW_PREDICT_FALSE(!ParseUnsigned(s + 2, 2, &minutes))) {
return false;
}
if (ARROW_PREDICT_FALSE(!ParseUnsigned(s + 4, 2, &seconds))) {
return false;
}
if (ARROW_PREDICT_FALSE(hours >= 24)) {
return false;
}
if (ARROW_PREDICT_FALSE(minutes >= 60)) {
return false;
}
if (ARROW_PREDICT_FALSE(seconds >= 60)) {
return false;
}
*out = std::chrono::duration_cast<Duration>(std::chrono::hours(hours) +
std::chrono::minutes(minutes) +
std::chrono::seconds(seconds));
return true;
}

static inline bool ParseSubSeconds(const char* s, size_t length, TimeUnit::type unit,
uint32_t* out) {
// The decimal point has been peeled off at this point
Expand Down Expand Up @@ -669,7 +698,7 @@ static inline bool ParseTimestampISO8601(const char* s, size_t length,
// We allow the following zone offset formats:
// - (none)
// - Z
// - [+-]HH(:?MM)?
// - [+-]HH(:?MM(:?SS)?)?
//
// We allow the following formats for all units:
// - "YYYY-MM-DD"
Expand Down Expand Up @@ -712,30 +741,48 @@ static inline bool ParseTimestampISO8601(const char* s, size_t length,
if (s[length - 1] == 'Z') {
--length;
if (out_zone_offset_present) *out_zone_offset_present = true;
} else if (s[length - 3] == '+' || s[length - 3] == '-') {
// [+-]HH
length -= 3;
if (ARROW_PREDICT_FALSE(!detail::ParseHH(s + length + 1, &zone_offset))) {
} else if (length >= 19 && (s[length - 9] == '+' || s[length - 9] == '-') &&
(s[length - 6] == ':') && (s[length - 3] == ':')) {
// [+-]HH:MM:SS
length -= 9;
if (ARROW_PREDICT_FALSE(!detail::ParseHH_MM_SS(s + length + 1, &zone_offset))) {
return false;
}
if (s[length] == '+') zone_offset *= -1;
if (out_zone_offset_present) *out_zone_offset_present = true;
} else if (s[length - 5] == '+' || s[length - 5] == '-') {
// [+-]HHMM
length -= 5;
if (ARROW_PREDICT_FALSE(!detail::ParseHHMM(s + length + 1, &zone_offset))) {
} else if (length >= 17 && (s[length - 7] == '+' || s[length - 7] == '-')) {
// [+-]HHMMSS
length -= 7;
if (ARROW_PREDICT_FALSE(!detail::ParseHHMMSS(s + length + 1, &zone_offset))) {
return false;
}
if (s[length] == '+') zone_offset *= -1;
if (out_zone_offset_present) *out_zone_offset_present = true;
} else if ((s[length - 6] == '+' || s[length - 6] == '-') && (s[length - 3] == ':')) {
} else if (length >= 16 && (s[length - 6] == '+' || s[length - 6] == '-') &&
(s[length - 3] == ':')) {
// [+-]HH:MM
length -= 6;
if (ARROW_PREDICT_FALSE(!detail::ParseHH_MM(s + length + 1, &zone_offset))) {
return false;
}
if (s[length] == '+') zone_offset *= -1;
if (out_zone_offset_present) *out_zone_offset_present = true;
} else if (length >= 15 && (s[length - 5] == '+' || s[length - 5] == '-')) {
// [+-]HHMM
length -= 5;
if (ARROW_PREDICT_FALSE(!detail::ParseHHMM(s + length + 1, &zone_offset))) {
return false;
}
if (s[length] == '+') zone_offset *= -1;
if (out_zone_offset_present) *out_zone_offset_present = true;
} else if (length >= 13 && (s[length - 3] == '+' || s[length - 3] == '-')) {
// [+-]HH
length -= 3;
if (ARROW_PREDICT_FALSE(!detail::ParseHH(s + length + 1, &zone_offset))) {
return false;
}
if (s[length] == '+') zone_offset *= -1;
if (out_zone_offset_present) *out_zone_offset_present = true;
}

seconds_type seconds_since_midnight;
Expand Down
8 changes: 8 additions & 0 deletions cpp/src/arrow/util/value_parsing_test.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -722,6 +722,10 @@ TEST(StringConversion, ToTimestampDateTime_ISO8601) {
AssertConversionFails(type, "1970-01-01 00:00:00-00000");
AssertConversionFails(type, "1970-01-01 00:00:00+2400");
AssertConversionFails(type, "1970-01-01 00:00:00+00:99");
AssertConversionFails(type, "1970-01-01 00:00:00+00:00:60");
AssertConversionFails(type, "1970-01-01 00:00:00+000060");
AssertConversionFails(type, "1970-01-01 00:00:00+240000");
AssertConversionFails(type, "1970-01-01 00:00:00+24:00:00");
}
{
TimestampType type{TimeUnit::MILLI};
Expand All @@ -742,6 +746,10 @@ TEST(StringConversion, ToTimestampDateTime_ISO8601) {
AssertConversion(type, "2018-11-13 17:11:10.123-01", 1542129070123LL + 3600000LL);
AssertConversion(type, "2018-11-13 17:11:10.123-0117", 1542129070123LL + 4620000LL);
AssertConversion(type, "2018-11-13 17:11:10.123-01:17", 1542129070123LL + 4620000LL);
AssertConversion(type, "2018-11-13 17:11:10.123+004430", 1542129070123LL - 2670000LL);
AssertConversion(type, "2018-11-13 17:11:10.123+00:44:30", 1542129070123LL - 2670000LL);
AssertConversion(type, "2018-11-13 17:11:10.123-004430", 1542129070123LL + 2670000LL);
AssertConversion(type, "2018-11-13 17:11:10.123-00:44:30", 1542129070123LL + 2670000LL);

// Invalid subseconds
AssertConversionFails(type, "1900-02-28 12:34:56.1234");
Expand Down
15 changes: 12 additions & 3 deletions cpp/src/arrow/vendored/datetime/date.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6018,11 +6018,12 @@ to_stream(std::basic_ostream<CharT, Traits>& os, const CharT* fmt,
os.setstate(ios::failbit);
return os;
}
auto m = duration_cast<minutes>(*offset_sec);
auto neg = m < minutes{0};
m = date::abs(m);
auto abs_sec = date::abs(*offset_sec);
auto m = duration_cast<minutes>(abs_sec);
auto neg = *offset_sec < std::chrono::seconds{0};
auto h = duration_cast<hours>(m);
m -= h;
auto s = abs_sec - duration_cast<std::chrono::seconds>(duration_cast<minutes>(abs_sec));
if (neg)
os << CharT{'-'};
else
Expand All @@ -6035,6 +6036,14 @@ to_stream(std::basic_ostream<CharT, Traits>& os, const CharT* fmt,
if (m < minutes{10})
os << CharT{'0'};
os << m.count();
if (s != std::chrono::seconds{0})

@dmitry-chirkov-dremio dmitry-chirkov-dremio Oct 22, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What if a user calls Arrow’s strftime function with %z? The vendored date library defines %z as +/-hhmm and %Ez as +/-hh:mm, truncating offsets that include seconds. This change instead returns +000921 or +00:09:21. Please avoid changing the format contract and handle sub-minute offsets in the timestamp-to-string cast, for example by formatting these values in UTC as suggested in #54534.

{
if (modified != CharT{})
os << CharT{':'};
if (s < std::chrono::seconds{10})
os << CharT{'0'};
os << s.count();
}
command = nullptr;
modified = CharT{};
}
Expand Down
Loading