fix: calculate total seconds from interval fields for extract(epoch) (#19807)

## Which issue does this PR close?

<!--
We generally require a GitHub issue to be filed for all bug fixes and
enhancements and this helps us generate change logs for our releases.
You can link an issue to this PR using the GitHub syntax. For example
`Closes #123` indicates that this PR will close issue #123.
-->

- Closes #19799.

## Rationale for this change

<!--
Why are you proposing this change? If this is already explained clearly
in the issue then this section is not needed.
Explaining clearly why changes are proposed helps reviewers understand
your changes and offer better suggestions for fixes.
-->
The epoch function incorrectly used `date_part(array, DatePart::Second)`
for intervals which only extracts the seconds component of the interval
structure (0 for "15 minutes"), not the total seconds from all
components.

## What changes are included in this PR?

<!--
There is no need to duplicate the description in the issue here but it
is sometimes worth providing a summary of the individual changes in this
PR.
-->

- Modified epoch() function in date_part.rs to properly handle interval
types by extracting struct fields and calculating total seconds
  - Added regression tests in date_part.slt

## Are these changes tested?

<!--
We typically require tests for all PRs in order to:
1. Prevent the code from being accidentally broken by subsequent changes
2. Serve as another way to document the expected behavior of the code

If tests are not included in your PR, please explain why (for example,
are they covered by existing tests)?
-->

Yes. Added tests verifying correct conversion for all interval types and
precision levels.

## Are there any user-facing changes?

<!--
If there are user-facing changes then we may require documentation to be
updated before approving the PR.
-->
Yes. `extract(epoch from interval)` now returns correct total seconds
instead of 0.

<!--
If there are any breaking changes to public APIs, please add the `api
change` label.
-->

Author: lemorage

datafusion/functions/src/datetime/date_part.rs

@@ -26,13 +26,16 @@ use arrow::datatypes::DataType::{
     Date32, Date64, Duration, Interval, Time32, Time64, Timestamp,
 };
 use arrow::datatypes::TimeUnit::{Microsecond, Millisecond, Nanosecond, Second};
-use arrow::datatypes::{DataType, Field, FieldRef, TimeUnit};
+use arrow::datatypes::{
+    DataType, Field, FieldRef, IntervalUnit as ArrowIntervalUnit, TimeUnit,
+};
 use datafusion_common::types::{NativeType, logical_date};
 
 use datafusion_common::{
     Result, ScalarValue,
     cast::{
-        as_date32_array, as_date64_array, as_int32_array, as_time32_millisecond_array,
+        as_date32_array, as_date64_array, as_int32_array, as_interval_dt_array,
+        as_interval_mdn_array, as_interval_ym_array, as_time32_millisecond_array,
         as_time32_second_array, as_time64_microsecond_array, as_time64_nanosecond_array,
         as_timestamp_microsecond_array, as_timestamp_millisecond_array,
         as_timestamp_nanosecond_array, as_timestamp_second_array,
@@ -56,7 +59,7 @@ use datafusion_macros::user_doc;
     argument(
         name = "part",
         description = r#"Part of the date to return. The following date parts are supported:
-        
+
     - year
     - quarter (emits value in inclusive range [1, 4] based on which quartile of the year the date is in)
     - month
@@ -70,7 +73,7 @@ use datafusion_macros::user_doc;
     - nanosecond
     - dow (day of the week where Sunday is 0)
     - doy (day of the year)
-    - epoch (seconds since Unix epoch)
+    - epoch (seconds since Unix epoch for timestamps/dates, total seconds for intervals)
     - isodow (day of the week where Monday is 0)
 "#
     ),
@@ -349,6 +352,11 @@ fn seconds(array: &dyn Array, unit: TimeUnit) -> Result<ArrayRef> {
 
 fn epoch(array: &dyn Array) -> Result<ArrayRef> {
     const SECONDS_IN_A_DAY: f64 = 86400_f64;
+    // Note: Month-to-second conversion uses 30 days as an approximation.
+    // This matches PostgreSQL's behavior for interval epoch extraction,
+    // but does not represent exact calendar months (which vary 28-31 days).
+    // See: https://doxygen.postgresql.org/datatype_2timestamp_8h.html
+    const DAYS_PER_MONTH: f64 = 30_f64;
 
     let f: Float64Array = match array.data_type() {
         Timestamp(Second, _) => as_timestamp_second_array(array)?.unary(|x| x as f64),
@@ -373,7 +381,19 @@ fn epoch(array: &dyn Array) -> Result<ArrayRef> {
         Time64(Nanosecond) => {
             as_time64_nanosecond_array(array)?.unary(|x| x as f64 / 1_000_000_000_f64)
         }
-        Interval(_) | Duration(_) => return seconds(array, Second),
+        Interval(ArrowIntervalUnit::YearMonth) => as_interval_ym_array(array)?
+            .unary(|x| x as f64 * DAYS_PER_MONTH * SECONDS_IN_A_DAY),
+        Interval(ArrowIntervalUnit::DayTime) => as_interval_dt_array(array)?.unary(|x| {
+            x.days as f64 * SECONDS_IN_A_DAY + x.milliseconds as f64 / 1_000_f64
+        }),
+        Interval(ArrowIntervalUnit::MonthDayNano) => {
+            as_interval_mdn_array(array)?.unary(|x| {
+                x.months as f64 * DAYS_PER_MONTH * SECONDS_IN_A_DAY
+                    + x.days as f64 * SECONDS_IN_A_DAY
+                    + x.nanoseconds as f64 / 1_000_000_000_f64
+            })
+        }
+        Duration(_) => return seconds(array, Second),
         d => return exec_err!("Cannot convert {d:?} to epoch"),
     };
     Ok(Arc::new(f))

datafusion/sqllogictest/test_files/datetime/date_part.slt

@@ -936,6 +936,57 @@ SELECT extract(second from arrow_cast(NULL, 'Interval(MonthDayNano)'))
 ----
 NULL
 
+# extract epoch from intervals
+query R
+SELECT extract(epoch from interval '15 minutes')
+----
+900
+
+query R
+SELECT extract(epoch from interval '1 hour')
+----
+3600
+
+query R
+SELECT extract(epoch from interval '1 day')
+----
+86400
+
+query R
+SELECT extract(epoch from interval '1 month')
+----
+2592000
+
+query R
+SELECT extract(epoch from arrow_cast('3 days', 'Interval(DayTime)'))
+----
+259200
+
+query R
+SELECT extract(epoch from arrow_cast('100 milliseconds', 'Interval(MonthDayNano)'))
+----
+0.1
+
+query R
+SELECT extract(epoch from arrow_cast('500 microseconds', 'Interval(MonthDayNano)'))
+----
+0.0005
+
+query R
+SELECT extract(epoch from arrow_cast('2500 nanoseconds', 'Interval(MonthDayNano)'))
+----
+0.0000025
+
+query R
+SELECT extract(epoch from arrow_cast('1 month 2 days 500 milliseconds', 'Interval(MonthDayNano)'))
+----
+2764800.5
+
+query R
+SELECT extract(epoch from arrow_cast('2 months', 'Interval(YearMonth)'))
+----
+5184000
+
 statement ok
 create table t (id int, i interval) as values
   (0, interval '5 months 1 day 10 nanoseconds'),

docs/source/user-guide/sql/scalar_functions.md

@@ -2531,7 +2531,7 @@ date_part(part, expression)
   - nanosecond
   - dow (day of the week where Sunday is 0)
   - doy (day of the year)
-  - epoch (seconds since Unix epoch)
+  - epoch (seconds since Unix epoch for timestamps/dates, total seconds for intervals)
   - isodow (day of the week where Monday is 0)
 
 - **expression**: Time expression to operate on. Can be a constant, column, or function.