IMPALA-13276: Revise the documentation of 'RUNTIME_FILTER_WAIT_TIME_MS'

This patch revises the documentation of the query option 'RUNTIME_FILTER_WAIT_TIME_MS' as well as the code comment for the same query option to make its meaning clearer. Change-Id: Ic98e23a902a65e4fa41a628d4a3edb1894660fb4 Reviewed-on: http://gerrit.cloudera.org:8080/21644 Tested-by: Impala Public Jenkins <[email protected]> Reviewed-by: Csaba Ringhofer <[email protected]>
apache · Aug 12, 2024 · 6736063 · 6736063
1 parent b02e3c8
commit 6736063
Show file tree

Hide file tree

Showing 3 changed files with 17 additions and 8 deletions.
diff --git a/common/thrift/Query.thrift b/common/thrift/Query.thrift
@@ -238,8 +238,10 @@ struct TQueryOptions {
   // be rounded up to the nearest power of two.
   38: optional i32 runtime_bloom_filter_size = 1048576
 
-  // Time in ms to wait until runtime filters are delivered. If 0, the default defined
-  // by the startup flag of the same name is used.
+  // Time in ms to wait until runtime filters are delivered. Note that the wait time for
+  // a runtime filter is with respect to the start of processing the query in the given
+  // executor instead of the beginning of the Open phase of a scan node. If 0, the
+  // default defined by the startup flag of the same name is used.
   39: optional i32 runtime_filter_wait_time_ms = 0
 
   // If true, per-row runtime filtering is disabled

diff --git a/docs/topics/impala_runtime_filter_wait_time_ms.xml b/docs/topics/impala_runtime_filter_wait_time_ms.xml
@@ -40,6 +40,13 @@ under the License.
       adjusts the settings for the runtime filtering feature.
       It specifies a time in milliseconds that each scan node waits for
       runtime filters to be produced by other plan fragments.
+      Note that the wait time for a runtime filter is with respect to the start of
+      processing the query in the given executor instead of the beginning of the Open
+      phase of a scan node. For instance, a scan node could start so late that at the
+      beginning of the Open phase of the scan node, the amount of time passed since the
+      start of query processing was already greater than the value of
+      <codeph>RUNTIME_FILTER_WAIT_TIME_MS</codeph>. In such a case, even though the
+      runtime filter has not arrived yet, the scan node will not wait any longer.
     </p>
 
     <p conref="../shared/impala_common.xml#common/type_integer"/>

diff --git a/docs/topics/impala_runtime_filtering.xml b/docs/topics/impala_runtime_filtering.xml
@@ -242,12 +242,12 @@ under the License.
         <codeph>RUNTIME_FILTER_WAIT_TIME_MS</codeph> query option.
       </p>
       <p>
-        By default, each scan node waits for up to 1 second (1000 milliseconds)
-        for filters to arrive. If all filters have not arrived within the
-        specified interval, the scan node proceeds, using whatever filters
-        did arrive to help avoid reading unnecessary data. If a filter arrives
-        after the scan node begins reading data, the scan node applies that
-        filter to the data that is read after the filter arrives, but not to
+        The time is counted from the start of executing the query — see the query
+        option's doc page for details.
+        If all filters have not arrived within the specified interval, the scan node
+        proceeds, using whatever filters did arrive to help avoid reading unnecessary
+        data. If a filter arrives after the scan node begins reading data, the scan node
+        applies that filter to the data that is read after the filter arrives, but not to
         the data that was already read.
       </p>
       <p>