Added --unmapped-val <val> option to bedmap; fixed named-pipe bug whe…

…n one file is passed to bedmap.

applications/bed/bedmap/src/Input.hpp

@@ -79,7 +79,7 @@ namespace BedMap {
         precision_(6), useScientific_(false), useMinMemory_(false), setPrec_(false), numFiles_(0),
         minRefFields_(0), minMapFields_(0), errorCheck_(false), sweepAll_(false),
         outDelim_("|"), multiDelim_(";"), fastMode_(false), rangeAlias_(false),
-        chrom_("all"), skipUnmappedRows_(false) {
+        chrom_("all"), skipUnmappedRows_(false), unmappedVal_("") {
 
       // Process user's operation options
       if ( argc <= 1 )
@@ -108,6 +108,12 @@ namespace BedMap {
           fastMode_ = true;
         } else if ( next == "sweep-all" ) { // --> sweep through all of second file
           sweepAll_ = true;
+        } else if ( next == "unmapped-val" ) {
+          Ext::Assert<ArgError>(unmappedVal_.empty(), "--unmapped-val specified multiple times");
+          Ext::Assert<ArgError>(argcntr < argc, "No value given for --unmapped-val");
+          unmappedVal_ = argv[argcntr++];
+          Ext::Assert<ArgError>(unmappedVal_.find("--") != 0,
+                                "Apparent option: " + std::string(argv[argcntr]) + " where <val> expected for --unmapped-val.");
         } else if ( next == "delim" ) {
           Ext::Assert<ArgError>(outDelim_ == "|", "--delim specified multiple times");
           Ext::Assert<ArgError>(argcntr < argc, "No output delimiter given");
@@ -399,6 +405,7 @@ namespace BedMap {
     bool rangeAlias_;
     std::string chrom_;
     bool skipUnmappedRows_;
+    std::string unmappedVal_;
 
   private:
     struct MapFields {
@@ -480,6 +487,7 @@ namespace BedMap {
     usage << "      --sci                 Use scientific notation for score outputs.                              \n";
     usage << "      --skip-unmapped       Print no output for a row with no mapped elements.                      \n";
     usage << "      --sweep-all           Ensure <map-file> is read completely (helps to prevent broken pipes).   \n";
+    usage << "      --unmapped-val <val>  Use <val> in place of the empty string on unmapped --echo-map* ops.     \n";
     usage << "      --version             Print program information.                                              \n";
     usage << "                                                                                                    \n";
     usage << "                                                                                                    \n";

docs/content/reference/statistics/bedmap.rst

@@ -82,6 +82,7 @@ The ``--help`` option describes the various mapping and analytical operations an
         --sci                 Use scientific notation for score outputs.
         --skip-unmapped       Print no output for a row with no mapped elements.
         --sweep-all           Ensure <map-file> is read completely (helps to prevent broken pipes).
+        --unmapped-val <val>  Use <val> in place of the empty string on unmapped --echo-map* ops.
         --version             Print program information.
 
       Overlap Options (At most, one may be selected.  By default, --bp-ovr 1 is used):

docs/content/revision-history.rst

@@ -23,6 +23,8 @@ Released: **TBD**
 
   * To avoid mapping problems, map elements should not contain spaces in the ID or subsequent non-interval fields. Use of the :code:`--ec` can help identify problems in map input, at the cost of a longer runtime. The documentation is clarified to warn users about avoiding spaces in map input. Thanks to Wouter Meuleman for the report and providing test input.
 
+  * Added --unmapped-val <val> option, where <val> replaces the empty string output of --echo-map* operations when there are no mapped elements.  --min/max-element operators will give results as before (the empty string).
+
 =================
 Previous versions
 =================

interfaces/general-headers/algorithm/visitors/helpers/ProcessBedVisitorRow.hpp

@@ -378,15 +378,18 @@ namespace Visitors {
     struct PrintUniqueRangeIDs : private Visitors::Helpers::PrintDelim {
       typedef Visitors::Helpers::PrintDelim Base;
 
-      explicit PrintUniqueRangeIDs(const std::string& delim = ";")
-          : Base(delim)
+      explicit PrintUniqueRangeIDs(const std::string& delim = ";", const std::string& onEmpty = "")
+          : Base(delim), oe_(onEmpty)
         { /* */ }
 
       template <typename Iter>
       void operator()(Iter beg, Iter end) const {
         typedef std::set<std::string> SortType;
-        if ( beg == end )
+        if ( beg == end ) {
+          if ( !oe_.empty() )
+            PrintTypes::Print(oe_.c_str());
           return;
+        }
         SortType srt;
         while ( beg != end ) {
           srt.insert((*beg)->id());
@@ -400,6 +403,9 @@ namespace Visitors {
           PrintTypes::Print(i->c_str());
         } // while
       }
+
+    private:
+      std::string oe_;
     };
 
     //=========================
@@ -448,18 +454,22 @@ namespace Visitors {
     //=====================
     template <typename PrintType>
     struct PrintGenomicRange {
+      typedef PrintType PType;
 
-      explicit PrintGenomicRange(const PrintType& p = PrintType())
-          : pt_(p)
+      explicit PrintGenomicRange(const PrintType& p = PrintType(), const std::string& onEmpty = "")
+          : pt_(p), oe_(onEmpty)
         { /* */ }
 
       template <typename Iter>
       void operator()(Iter beg, Iter end) const {
         // It is often possible/likely that beg->end is not in an order
         //  of the original input file (sorted by sort-bed), due to
         //  issues of BED types (see BedBaseVisitor.hpp).
-        if ( beg == end )
+        if ( beg == end ) {
+          if ( !oe_.empty() )
+            PrintTypes::Print(oe_.c_str());
           return;
+        }
 
         typename std::remove_const<typename std::remove_pointer<typename Iter::value_type>::type>::type val = **beg;
         while ( ++beg != end ) {
@@ -474,6 +484,7 @@ namespace Visitors {
 
     private:
       PrintType pt_;
+      std::string oe_;
     };
 
   } // namespace BedHelpers

interfaces/general-headers/algorithm/visitors/helpers/ProcessVisitorRow.hpp

@@ -150,14 +150,18 @@ namespace Visitors {
       typedef PrintType PType;
 
       explicit PrintRangeDelim(const PrintType& p = PrintType(),
-                               const std::string& delim = ";")
-          : Base(delim), pt_(p)
+                               const std::string& delim = ";",
+                               const std::string& onEmpty = "")
+          : Base(delim), pt_(p), oe_(onEmpty)
         { /* */ }
 
       template <typename Iter>
       void operator()(Iter beg, Iter end) const {
-        if ( beg == end )
+        if ( beg == end ) {
+          if ( !oe_.empty() )
+            PrintTypes::Print(oe_.c_str());
           return;
+        }
         pt_.operator()(*beg);
         while ( ++beg != end ) {
           Base::operator()();
@@ -167,6 +171,7 @@ namespace Visitors {
 
     private:
       PrintType pt_;
+      std::string oe_;
     };
 
   } // namespace Helpers