hpcc-systems · jeclrsg · Jan 5, 2026 · Jan 6, 2026 · Dec 4, 2025 · Jan 7, 2026
diff --git a/common/eventconsumption/eventindexhotspot.cpp b/common/eventconsumption/eventindexhotspot.cpp
@@ -35,10 +35,17 @@ class CHotspotEventVisitor : public CInterfaceOf<IEventVisitor>
         // The lowest and highest bucket number with observed activity.
         using Range = std::pair<bucket_type, bucket_type>;
         // The observed activity for a single bucket kind.
-        struct Activity
+        class Activity
         {
+        public:
             Buckets buckets;
             Range range{std::numeric_limits<bucket_type>::max(), 0};
+
+            Activity() = default;
+            bool hasActivity() const
+            {
+                return !buckets.empty();
+            }
         };
     public:
         CActivity(CHotspotEventVisitor& _container, __uint64 _id)
@@ -62,11 +69,13 @@ class CHotspotEventVisitor : public CInterfaceOf<IEventVisitor>
         void forEachBucket(IBucketVisitor& visitor)
         {
             const char* path = container.operation.queryMetaInfoState().queryFilePath(id);
+            if (!hasActivity()) // suppress inactive files
+                return;
             visitor.arrive(id, path);
             for (unsigned kind = 0; kind < NumNodeKinds; kind++)
             {
                 Activity& act = activity[kind];
-                if (act.range.first != std::numeric_limits<bucket_type>::max() && !act.buckets.empty())
+                if (act.hasActivity())
                 {
                     for (bucket_type bucket = act.range.first; bucket <= act.range.second; bucket++)
                     {
@@ -102,6 +111,16 @@ class CHotspotEventVisitor : public CInterfaceOf<IEventVisitor>
             return 0;
         }
 
+        bool hasActivity() const
+        {
+            for (const Activity& act : activity)
+            {
+                if (act.hasActivity())
+                    return true;
+            }
+            return false;
+        }
+
     private:
         CHotspotEventVisitor& container; // back reference to obtain shared configuration
         __uint64 id{0};

diff --git a/common/eventconsumption/eventindexsummarize.cpp b/common/eventconsumption/eventindexsummarize.cpp
@@ -665,6 +665,11 @@ class CTraceCollector : public CSummaryCollector
 public: // IEventVisitor
     virtual bool visitEvent(CEvent& event) override
     {
+        // Allow EventQueryStart events to pass through; trace ID to service name
+        // mapping is built earlier by the metadata parser (CMetaInfoState::visitEvent)
+        if (event.queryType() == EventQueryStart)
+            return true;
+
         // Implicit event filter applied unconditionally
         if (queryEventContext(event.queryType()) != EventCtxIndex)
             return true;
@@ -688,7 +693,7 @@ class CTraceCollector : public CSummaryCollector
     virtual void summarize() override
     {
         // Determine which node kinds have data
-        bool haveNodeKindEntries[NumNodeKinds] = {false,};
+        std::fill_n(haveNodeKindEntries, NumNodeKinds, false);
         for (Cache::value_type& entry : stats)
         {
             TraceStats& traceStats = entry.second;
@@ -701,23 +706,18 @@ class CTraceCollector : public CSummaryCollector
                 }
             }
         }
+        if (summarization == IndexSummarization::byService)
+            summarizeByService();
+        else
+            summarizeByTrace();
+    }
 
+protected:
+    void summarizeByTrace()
+    {
         StringBuffer line;
         appendCSVColumns(line, "Trace ID", "Service Name", "First Timestamp", "Last Timestamp", "Unique Files");
-
-        // Add aggregate statistics headers
-        appendNodeKindHeaders(line, "Total ");
-
-        // Add node kind specific headers only for kinds that have data
-        for (unsigned nodeKind = 0; nodeKind < NumNodeKinds; nodeKind++)
-        {
-            if (!haveNodeKindEntries[nodeKind])
-                continue;
-            const char* nodeKindName = mapNodeKind((NodeKind)nodeKind);
-            VStringBuffer prefix("%c%s ", toupper(*nodeKindName), nodeKindName + 1);
-            appendNodeKindHeaders(line, prefix);
-        }
-
+        appendNodeKindHeaders(line);
         outputLine(line);
 
         for (Cache::value_type& entry : stats)
@@ -749,7 +749,6 @@ class CTraceCollector : public CSummaryCollector
 
             // Create aggregate statistics across all node kinds
             NodeKindStats aggregateStats = {};
-            std::unordered_map<IndexHashKey, uint32_t, IndexHashKeyHash> allNodeMemory;
 
             for (unsigned nodeKind = 0; nodeKind < NumNodeKinds; nodeKind++)
                 aggregateStats.addStats(traceStats.kinds[nodeKind]);
@@ -770,7 +769,115 @@ class CTraceCollector : public CSummaryCollector
         }
     }
 
+    void summarizeByService()
+    {
+        // Extended TraceStats with trace counter for service summarization
+        class ServiceStats : public TraceStats
+        {
+        public:
+            uint32_t traceCount{0}; // Count of traces represented in this service
+            NodeKindStats totals;
+
+            ServiceStats() = default;
+            void addStats(const TraceStats& other)
+            {
+                for (unsigned nodeKind = 0; nodeKind < NumNodeKinds; nodeKind++)
+                {
+                    const NodeKindStats& nodeKindStats = other.kinds[nodeKind];
+                    kinds[nodeKind].addStats(nodeKindStats);
+                    totals.addStats(nodeKindStats);
+                }
+            }
+        };
+
+        // Aggregate trace stats by service name
+        std::unordered_map<std::string, ServiceStats> serviceStats;
+
+        for (Cache::value_type& entry : stats)
+        {
+            const TraceHashKey& traceKey = entry.first;
+            TraceStats& traceStats = entry.second;
+
+            // Find service name for this trace ID, or use empty string if not found
+            const char* serviceName = operation.queryMetaInfoState().queryServiceName(traceKey.traceId.str());
+            if (!serviceName)
+                serviceName = "";
+            ServiceStats& serviceAggregateStats = serviceStats[serviceName];
+
+            // Aggregate the stats
+            serviceAggregateStats.addStats(traceStats);
+
+            // Increment trace count
+            serviceAggregateStats.traceCount++;
+
+            // Merge unique files
+            serviceAggregateStats.uniqueFileIds.insert(traceStats.uniqueFileIds.begin(), traceStats.uniqueFileIds.end());
+
+            // Update timestamp ranges
+            if (traceStats.firstTimestamp != UINT64_MAX)
+                serviceAggregateStats.firstTimestamp = std::min(serviceAggregateStats.firstTimestamp, traceStats.firstTimestamp);
+            if (traceStats.lastTimestamp > 0)
+                serviceAggregateStats.lastTimestamp = std::max(serviceAggregateStats.lastTimestamp, traceStats.lastTimestamp);
+        }
+
+        // Output aggregated service stats
+        StringBuffer line;
+        appendCSVColumns(line, "Service Name", "Trace Count", "First Timestamp", "Last Timestamp", "Unique Files");
+        appendNodeKindHeaders(line);
+        outputLine(line);
+
+        for (auto& serviceEntry : serviceStats)
+        {
+            const std::string& serviceName = serviceEntry.first;
+            ServiceStats& serviceStatsEntry = serviceEntry.second;
+
+            // Format timestamps using CDateTime (empty if no valid timestamps were found)
+            StringBuffer firstTsStr, lastTsStr;
+            if (serviceStatsEntry.firstTimestamp != UINT64_MAX)
+            {
+                CDateTime firstDt;
+                firstDt.setTimeStampNs(serviceStatsEntry.firstTimestamp);
+                firstDt.getString(firstTsStr);
+            }
+            if (serviceStatsEntry.lastTimestamp > 0)
+            {
+                CDateTime lastDt;
+                lastDt.setTimeStampNs(serviceStatsEntry.lastTimestamp);
+                lastDt.getString(lastTsStr);
+            }
+
+            const char* displayServiceName = serviceName.empty() ? "<no service>" : serviceName.c_str();
+            appendCSVColumns(line, displayServiceName, serviceStatsEntry.traceCount, firstTsStr.str(), lastTsStr.str(), serviceStatsEntry.uniqueFileIds.size());
+            appendNodeKindData(line, serviceStatsEntry.totals);
+            for (unsigned nodeKind = 0; nodeKind < NumNodeKinds; nodeKind++)
+            {
+                if (!haveNodeKindEntries[nodeKind])
+                    continue;
+                const NodeKindStats& nodeStats = serviceStatsEntry.kinds[nodeKind];
+                appendNodeKindData(line, nodeStats);
+            }
+            outputLine(line);
+        }
+    }
+
 protected:
+
+    void appendNodeKindHeaders(StringBuffer& line)
+    {
+        // Add aggregate statistics headers
+        appendNodeKindHeaders(line, "Total ");
+
+        // Add node kind specific headers only for kinds that have data
+        for (unsigned nodeKind = 0; nodeKind < NumNodeKinds; nodeKind++)
+        {
+            if (!haveNodeKindEntries[nodeKind])
+                continue;
+            const char* nodeKindName = mapNodeKind((NodeKind)nodeKind);
+            VStringBuffer prefix("%c%s ", toupper(*nodeKindName), nodeKindName + 1);
+            appendNodeKindHeaders(line, prefix);
+        }
+    }
+
     void appendNodeKindHeaders(StringBuffer& line, const char* prefix)
     {
         appendCSVEventsHeaders(line, prefix);
@@ -807,6 +914,8 @@ class CTraceCollector : public CSummaryCollector
 protected:
     using Cache = std::unordered_map<TraceHashKey, TraceStats, TraceHashKeyHash>;
     Cache stats;
+    bool haveNodeKindEntries[NumNodeKinds] = {false,};
+
 };
 
 bool CIndexFileSummary::doOp()
@@ -820,6 +929,7 @@ bool CIndexFileSummary::doOp()
         collector.setown(new CNodeCollector(*this, summarization, out));
         break;
     case IndexSummarization::byTrace:
+    case IndexSummarization::byService:
         collector.setown(new CTraceCollector(*this, summarization, out));
         break;
     default:

diff --git a/common/eventconsumption/eventindexsummarize.h b/common/eventconsumption/eventindexsummarize.h
@@ -30,6 +30,7 @@ enum class IndexSummarization
     byNodeKind,
     byNode,
     byTrace,
+    byService,
 };
 
 class event_decl CIndexFileSummary : public CEventConsumingOp

diff --git a/docs/EN_US/ECLLanguageReference/ECLR_mods/BltInFunc-TOJSON.xml b/docs/EN_US/ECLLanguageReference/ECLR_mods/BltInFunc-TOJSON.xml
@@ -4,16 +4,16 @@
 <sect1 id="TOJSON">
   <title>TOJSON</title>
 
-  <para><emphasis role="bold">TOJSON<indexterm>
+  <para><emphasis role="bold">TOJSON(</emphasis><emphasis> record
+  </emphasis><emphasis role="bold">[,ALL])</emphasis><indexterm>
       <primary>TOJSON</primary>
     </indexterm><indexterm>
       <primary>TOJSON function</primary>
-    </indexterm>(</emphasis><emphasis> record </emphasis><emphasis
-  role="bold">)</emphasis></para>
+    </indexterm></para>
 
   <para><informaltable colsep="1" frame="all" rowsep="1">
       <tgroup cols="2">
-        <colspec colwidth="77.95pt" />
+        <colspec colwidth="77.95pt"/>
 
         <tbody>
           <row>
@@ -22,6 +22,15 @@
             <entry>The row (record) of data to convert to JSON format.</entry>
           </row>
 
+          <row>
+            <entry><emphasis role="bold">ALL</emphasis></entry>
+
+            <entry>Specifies fields with empty strings are included in the
+            output. Without the ALL flag, <emphasis
+            role="bold">TOJSON</emphasis> omits fields that contain only
+            whitespace or are empty after trimming.</entry>
+          </row>
+
           <row>
             <entry>Return:</entry>
 

diff --git a/docs/EN_US/ECLLanguageReference/ECLR_mods/BltInFunc-TOXML.xml b/docs/EN_US/ECLLanguageReference/ECLR_mods/BltInFunc-TOXML.xml
@@ -4,16 +4,16 @@
 <sect1 id="TOXML">
   <title>TOXML</title>
 
-  <para><emphasis role="bold">TOXML<indexterm>
+  <para><emphasis role="bold">TOXML(</emphasis><emphasis> record
+  </emphasis><emphasis role="bold">[,ALL])</emphasis><indexterm>
       <primary>TOXML</primary>
     </indexterm><indexterm>
       <primary>TOXML function</primary>
-    </indexterm>(</emphasis><emphasis> record </emphasis><emphasis
-  role="bold">)</emphasis></para>
+    </indexterm></para>
 
   <para><informaltable colsep="1" frame="all" rowsep="1">
       <tgroup cols="2">
-        <colspec colwidth="77.95pt" />
+        <colspec colwidth="77.95pt"/>
 
         <tbody>
           <row>
@@ -23,6 +23,15 @@
             format.</entry>
           </row>
 
+          <row>
+            <entry><emphasis role="bold">ALL</emphasis></entry>
+
+            <entry>Specifies fields with empty strings are included in the
+            output. Without the ALL flag, <emphasis
+            role="bold">TOXML</emphasis> omits fields that contain only
+            whitespace or are empty after trimming.</entry>
+          </row>
+
           <row>
             <entry>Return:</entry>
 
@@ -32,8 +41,6 @@
       </tgroup>
     </informaltable></para>
 
-  <para></para>
-
   <para>The <emphasis role="bold">TOXML </emphasis>function returns a single
   UTF-8 string with the data in the <emphasis>record</emphasis> re-formatted
   as XML. If the RECORD structure of the <emphasis>record</emphasis> has

diff --git a/docs/common/Version.xml b/docs/common/Version.xml
@@ -5,11 +5,11 @@
   <chapterinfo>
     <date id="DateVer">DEVELOPER NON-GENERATED VERSION</date>
 
-    <releaseinfo id="FooterInfo">© 2025 HPCC
+    <releaseinfo id="FooterInfo">© 2026 HPCC
     Systems<superscript>®</superscript>. All rights reserved</releaseinfo>
 
     <copyright id="Copyright">
-      <year>2025 HPCC Systems<superscript>®</superscript>. All rights
+      <year>2026 HPCC Systems<superscript>®</superscript>. All rights
       reserved</year>
     </copyright>
   </chapterinfo>
@@ -23,9 +23,9 @@
     serve one purpose and that is to store the chapterinfo and other information
     used by several document components.</para>
 
-    <para id="CHMVer">2025 Version ${DOC_VERSION}</para>
+    <para id="CHMVer">2026 Version ${DOC_VERSION}</para>
 
-    <remark id="YrVer">2025</remark>
+    <remark id="YrVer">2026</remark>
 
     <para>The following line is the code to be put into the document you wish
     to include the above version info in:</para>

diff --git a/docs/common/Version.xml.in b/docs/common/Version.xml.in
@@ -3,13 +3,13 @@
 "http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd">
 <chapter>
   <chapterinfo>
-    <date id="DateVer">2025 Version ${DOC_VERSION}</date>
+    <date id="DateVer">2026 Version ${DOC_VERSION}</date>
 
-    <releaseinfo id="FooterInfo">© 2025 HPCC
+    <releaseinfo id="FooterInfo">© 2026 HPCC
     Systems<superscript>®</superscript>. All rights reserved</releaseinfo>
 
     <copyright id="Copyright">
-      <year>2025 HPCC Systems<superscript>®</superscript>. All rights
+      <year>2026 HPCC Systems<superscript>®</superscript>. All rights
       reserved</year>
     </copyright>
   </chapterinfo>
@@ -24,9 +24,9 @@
     serve one purpose and that is to store the chapterinfo the above sections
     that are being used by several other documents.</para>
 
-    <para id="CHMVer">2025 Version ${DOC_VERSION}</para>
+    <para id="CHMVer">2026 Version ${DOC_VERSION}</para>
 
-    <remark id="YrVer">2025</remark>
+    <remark id="YrVer">2026</remark>
 
     <para>The following line is the code to be put into the document you wish
     to include the above version info in:</para>