Added distinct_sort_num option to VectorOps for groupBy.

src/groupBy/groupBy.cpp

@@ -83,8 +83,7 @@ int groupby_main(int argc, char* argv[]) {
     bool InputHaveHeaderLine = false;
     bool ignoreCase    = false;
     int precision = 21;
-
-    // check to see if we should print out some help
+     // check to see if we should print out some help
     if(argc <= 1) showHelp = true;
 
     for(int i = 1; i < argc; i++) {
@@ -186,6 +185,7 @@ int groupby_main(int argc, char* argv[]) {
                 i++;
             }
         }
+
         else {
             cerr << endl 
                  << "*****ERROR: Unrecognized parameter: " 
@@ -216,7 +216,7 @@ int groupby_main(int argc, char* argv[]) {
             (ops[i] != "antimode") && (ops[i] != "stdev") &&
             (ops[i] != "sstdev") && (ops[i] != "count") && 
             (ops[i] != "count_distinct") && (ops[i] != "collapse") && 
-            (ops[i] != "distinct") && (ops[i] != "concat") && 
+            (ops[i] != "distinct") && (ops[i] != "distinct_sort_num") && (ops[i] != "concat") &&
             (ops[i] != "freqdesc") && (ops[i] != "freqasc") &&
             (ops[i] != "first") && (ops[i] != "last") ) 
         {
@@ -318,6 +318,7 @@ void groupby_help(void) {
     cerr                         << "\t\t\t    stdev, sstdev (sample standard dev.)," << endl;
     cerr                         << "\t\t\t    collapse (i.e., print a comma separated list (duplicates allowed)), " << endl;
     cerr                         << "\t\t\t    distinct (i.e., print a comma separated list (NO duplicates allowed)), " << endl;
+    cerr                         << "\t\t\t    distinct_sort_num (as distinct, but sorted numerically), " << endl;
     cerr                         << "\t\t\t    concat   (i.e., merge values into a single, non-delimited string), " << endl;
     cerr                         << "\t\t\t    freqdesc (i.e., print desc. list of values:freq)" << endl;
     cerr                         << "\t\t\t    freqasc (i.e., print asc. list of values:freq)" << endl;
@@ -473,7 +474,7 @@ void GroupBy (const string &inFile,
 void ReportSummary(const vector<string> &group, 
                    const vector<vector<string> > &data, 
                    const vector<string> &ops,
-                   int precision) 
+                   int precision)
 {
 
     vector<string> result;
@@ -497,6 +498,10 @@ void ReportSummary(const vector<string> &group,
         else if (op == "distinct") {
             result.push_back(vo.GetDistinct());
         }
+        else if (op == "distinct_sort_num") {
+            result.push_back(vo.GetDistinctSortNum());
+        }
+
         else if (op == "concat") {
             result.push_back(vo.GetConcat());
         }

src/utils/VectorOps/VectorOps.cpp

@@ -267,6 +267,25 @@ string VectorOps::GetDistinct(void)
     return distinct.str();
 }
 
+string VectorOps::GetDistinctSortNum(void)
+{
+    ostringstream distinct;
+    // remove duplicate entries from the vector
+    // http://stackoverflow.com/questions/1041620/most-efficient-way-to-erase-duplicates-and-sort-a-c-vector
+    // convert the vec of strings to a vec of doubles
+    transform(_vecs.begin(), _vecs.end(), back_inserter(_vecd), MakeDouble);
+
+    sort( _vecd.begin(), _vecd.end() );
+    _vecd.erase( unique( _vecd.begin(), _vecd.end() ), _vecd.end() );
+
+    for( size_t i = 0; i < _vecd.size(); i++ ) {
+        if (i>0)
+            distinct << ",";
+        distinct << _vecd[i];
+    }
+    return distinct.str();
+}
+
 string VectorOps::GetFreqDesc(void)
 {
     // compute the frequency of each unique value
@@ -349,5 +368,3 @@ string VectorOps::GetLast(void)
 {
     return _vecs[_vecs.size() - 1];
 }
-
-

src/utils/VectorOps/VectorOps.h

@@ -72,7 +72,9 @@ public:
     // return a concatenation of all elements in the vector
     string GetConcat(void);           
     // return a comma-separated list of the _unique_ elements
-    string GetDistinct(void);         
+    string GetDistinct(void);
+    // return a comma-separated list of the _unique_ elements, sorted numerically
+    string GetDistinctSortNum(void);
     // return a histogram of values and their freqs. in desc. order of frequency
     string GetFreqDesc(void);         
     // return a histogram of values and their freqs. in asc. order of frequency