From 01ea2597d922774ff641d6a2eacae22979d68802 Mon Sep 17 00:00:00 2001
From: Joe Zhao <ztuowen@gmail.com>
Date: Fri, 22 May 2015 11:31:25 +0800
Subject: added shuffle option

---
 split.cpp                  | 14 +++++++++++---
 tools/fileDataProvider.cpp |  8 ++++----
 tools/fileDataProvider.h   |  9 +++++++++
 3 files changed, 24 insertions(+), 7 deletions(-)

diff --git a/split.cpp b/split.cpp
index be80545..e774ea9 100644
--- a/split.cpp
+++ b/split.cpp
@@ -43,6 +43,7 @@ int main(int argc, char **argv)
     desc.add_options()
             ("help,h", "produce help message")
             ("query,Q", "Query person count")
+            ("shuffle,s","shuffle input")
             ("count,c", po::value<int>(), "take number")
             ("take,a", po::value<string>(), "set output rid file 1(taken)")
             ("left,b", po::value<string>(), "set output rid file 2(left)")
@@ -68,9 +69,16 @@ int main(int argc, char **argv)
     vector<DataEntry*> a;
     vector<DataEntry*> b;
     dp.open();
-    dp.take(vm["count"].as<int>(),a,b);
-    outputRid(a,dp.getfSize(),vm["take"].as<string>());
-    outputRid(b,dp.getfSize(),vm["left"].as<string>());
+    if (vm.count("shuffle")) {
+        dp.getRidVector(a);
+        dp.shuffle(a);
+        outputRid(a, dp.getfSize(), vm["take"].as<string>());
+    }
+    else {
+        dp.take(vm["count"].as<int>(), a, b);
+        outputRid(a, dp.getfSize(), vm["take"].as<string>());
+        outputRid(b, dp.getfSize(), vm["left"].as<string>());
+    }
     dp.close();
     return 0;
 }
\ No newline at end of file
diff --git a/tools/fileDataProvider.cpp b/tools/fileDataProvider.cpp
index b1f3f5e..9be1132 100644
--- a/tools/fileDataProvider.cpp
+++ b/tools/fileDataProvider.cpp
@@ -124,7 +124,7 @@ int RidFileDP::getpSize() {
     return p.size();
 };
 
-void scrambler(vector<DataEntry*> &dat)
+void RidFileDP::shuffle(vector<DataEntry*> &dat)
 {
     DataEntry* e;
     int sz=(int)dat.size();
@@ -148,7 +148,7 @@ void RidFileDP::take(int n,vector<DataEntry*> &a,vector<DataEntry*> &b)
     a.clear();
     b.clear();
     std::vector<DataEntry*> &dat = d.getData();
-    scrambler(tmp);
+    shuffle(tmp);
     for (int i=0;i<dat.size();++i)
         tmp.push_back(dat[i]);
     int pos = 0;
@@ -170,8 +170,8 @@ void RidFileDP::take(int n,vector<DataEntry*> &a,vector<DataEntry*> &b)
     for (int i=0;i<tmp.size();++i)
         if (tmp[i]!=NULL)
             b.push_back(tmp[i]);
-    scrambler(a);
-    scrambler(b);
+    shuffle(a);
+    shuffle(b);
 }
 
 void RidFileDP::getAllDataSet(RidList &out){
diff --git a/tools/fileDataProvider.h b/tools/fileDataProvider.h
index 567c8e2..972a4c5 100644
--- a/tools/fileDataProvider.h
+++ b/tools/fileDataProvider.h
@@ -38,11 +38,20 @@ public:
     void readEntries();
     int getfSize() { if(!read) readEntries(); return d.getfSize();};
     int getpSize();
+    void shuffle(std::vector<DataEntry*> &dat);
     virtual void getAllDataSet(RidList &out);
     virtual int getDataSet(DataList &out);
     virtual int open(){fin.open(fname); eof=false;return 0;};
     virtual int close(){fin.close(); d.clear();return 0;};
     void take(int n,std::vector<DataEntry*> &a,std::vector<DataEntry*> &b);
+    void getRidVector(std::vector<DataEntry*> &rid){
+        if (!read)
+            readEntries();
+        rid.clear();
+        std::vector<DataEntry*> &dat = d.getData();
+        for (int i=0;i<dat.size();++i)
+            rid.push_back(dat[i]);
+    }
 };
 
 #endif
\ No newline at end of file
-- 
cgit v1.2.3-70-g09d2