Merge pull request #25 from brentp/groupby-columns

Groupby columns specified like cut (but retaining order)

src/groupBy/Makefile

@@ -16,7 +16,7 @@ INCLUDES = -I$(UTILITIES_DIR)/tabFile/ \
 # ----------------------------------
 # define our source and object files
 # ----------------------------------
-SOURCES= groupBy.cpp
+SOURCES= groupBy.cpp $(UTILITIES_DIR)/lineFileUtilities/lineFileUtilities.h
 OBJECTS= groupBy.o
 BUILT_OBJECTS= $(patsubst %,$(OBJ_DIR)/%,$(OBJECTS))
 

src/groupBy/groupBy.cpp

@@ -172,10 +172,10 @@ int groupby_main(int argc, char* argv[]) {
         // Split the column string sent by the user into discrete column numbers
         // A comma separated string is expected.
         vector<int> groupColumnsInt;
-        Tokenize(groupColumnsString, groupColumnsInt, ',');
+        TokenizeColumns(groupColumnsString, groupColumnsInt);
 
         vector<int> opColumnsInt;
-        Tokenize(opsColumnString, opColumnsInt, ',');
+        TokenizeColumns(opsColumnString, opColumnsInt);
 
         // sanity check the group columns
         for(size_t i = 0; i < groupColumnsInt.size(); ++i) {

src/utils/lineFileUtilities/lineFileUtilities.h

@@ -46,5 +46,40 @@ void Tokenize(const string &str, vector<int> &elems, char delimiter = '\t')
     }
 }
 
+// tokenize a column string into a list of integers.
+inline
+void TokenizeColumns(const string &str, vector<int> &elems) 
+{
+
+    // http://stackoverflow.com/questions/236129/how-to-split-a-string-in-c/236803#236803
+    // NOTE: this approach intentionally allows consecutive delimiters
+    vector<string> col_sets;
+    Tokenize(str, col_sets, ',');
+
+    for( size_t i = 0; i < col_sets.size(); i++ ) {
+        string col_set = col_sets[i];
+        if( string::npos == col_set.find("-") ){
+            elems.push_back(atoi(col_set.c_str()));
+        }
+        else {
+            vector<string> ends;
+            Tokenize(col_set, ends, '-');
+            int start = atoi(ends[0].c_str());
+            int end = atoi(ends[1].c_str());
+            if(start <= end){
+                for(int i = start; i <= end; i++){
+                    elems.push_back(i);
+                }
+            }
+            else {
+                for(int i = start; i >= end; i--){
+                    elems.push_back(i);
+                }
+            } 
+        }
+    }
+}
+
+
 #endif /* LINEFILEUTILITIES_H */
 

test/groupBy/test-groupby.sh

+lines_a=$(../../bin/groupBy -g 3-1 -o collapse -c 4 -i ../map/values3.bed | wc -l)
+lines_b=$(../../bin/groupBy -g 1-3 -o collapse -c 4 -i ../map/values3.bed | wc -l)
+lines_c=$(../../bin/groupBy -g 1,2,3 -o collapse -c 4 -i ../map/values3.bed | wc -l)
+lines_d=$(../../bin/groupBy -g 1-2,3 -o collapse -c 4 -i ../map/values3.bed | wc -l)
+
+check(){
+    if [ "$1" != "$2" ]; then
+        "fail groupby"
+    fi
+}
+
+check $lines_a $lines_b
+check $lines_a $lines_c
+check $lines_a $lines_d