diff --git a/MexFunc/IsWordInDic.cpp b/MexFunc/IsWordInDic.cpp
index 385e671..7c1cfe3 100644
--- a/MexFunc/IsWordInDic.cpp
+++ b/MexFunc/IsWordInDic.cpp
@@ -91,7 +91,7 @@ mxArray* writeToMatDouble(const double *data, int rowNum, int colNum) {
 输入：
 1. wd 文献摘要中的单词，二维cell
 2. dic 字典，按字母序排序的，一维cell
-3. threshold 保留超过阈值的列
+[3]. flagPrint
 输出：
 x 一维int(double)类型，表示在wd的每一行单词中，dic中是否有单词匹配上（匹配后，对应的坐标为设为1，否则为0），所有匹配的个数（统计每一行的匹配个数）
   dic每一个单词在wd所有行中出现的次数 
@@ -108,63 +108,56 @@ void mexFunction(int nlhs, mxArray* plhs[], int nrhs, const mxArray* prhs[]) {
 	Read2DWord(prhs[0], vvWd);
 	Read1DWord(prhs[1], vDic);
 	int rowNum = vvWd.size();
+	int colNum = vDic.size();
 
-	int threshold = 5;
+	int flagPrint = 0;
 	if (nrhs > 2) {
-		double* pThreshold = (double*)mxGetData(prhs[2]);
-		threshold = (int)pThreshold[0];
-		if (threshold < 5) threshold = 5;
+		double* pData = (double*)mxGetData(prhs[2]);
+		flagPrint = (int)pData[0];
 	}
 	finish = clock();
-	cout << "Load data time: " << (double)(finish - begin) / CLOCKS_PER_SEC << " s" << endl;
+	if (flagPrint == 2) cout << "Load data time: " << (double)(finish - begin) / CLOCKS_PER_SEC << " s" << endl;
 
-	vector<double> vXSum(vDic.size());
+	vector<double> vX(rowNum * colNum); // 一维表示二维
 
 	/* 统计dicr字典中，每个单词在wd中出现的次数 */
 	mid = clock();
-	unordered_map<string, int> umWordPos;
-	for (int i = 0; i < vDic.size(); ++i) umWordPos[vDic[i]] = i; // 记录单词位置
+	unordered_map<string, vector<int>> umWordPos;
+	for (int i = 0; i < vDic.size(); ++i) umWordPos[vDic[i]].push_back(i); // dic中可能存在重复,记录单词位置
 	unordered_set<int> usPos; // 多次出现在wd中的单词，只统计一次，这是原matlab代码的功能，是否需要修改？
 	vector<unordered_set<int>> vusX(rowNum); // 保存每一行中非零元的坐标
 	int row = 0;
+	// vector<double> vSum(colNum);
 	for (auto& vWd : vvWd) {
-		auto& usPos = vusX[row++];
+		auto& usPos = vusX[row];
 		for (auto& word : vWd) {
 			auto itr = umWordPos.find(word);
 			if (itr != umWordPos.end()) {
-				usPos.insert(itr->second);
+				for (auto pos : itr->second)
+					usPos.insert(pos);
 			}
 		}
 		for (auto idx : usPos) {
-			vXSum[idx] += 1;
+			vX[idx * rowNum + row] = 1; // matlab 列优先存储模式
+			// vSum[idx] += 1;
 		}
+		++row;
 	}
 	finish = clock();
-	cout << "Calc word occurrence time: " << (double)(finish - mid) / CLOCKS_PER_SEC << " s" << endl;
+	if (flagPrint == 2) cout << "Calc word occurrence time: " << (double)(finish - mid) / CLOCKS_PER_SEC << " s" << endl;
 
-	/* 计算xs */
-	mid = clock();
-	int colNum = 0;
-	vector<int> vColIdx;
-	for (int i = 0; i < vXSum.size(); ++i) {
-		if (vXSum[i] >= threshold) {
-			vColIdx.push_back(i);
-		}
-	}
-	colNum = vColIdx.size();
+	// for (auto& w : vDic) cout << umWordPos[w] << endl;
 
-	vector<double> vXsData(rowNum * colNum);
-
-	for (int i = 0; i < rowNum; ++i) {
-		for (int j = 0; j < colNum; ++j) {
-			if (vusX[i].find(vColIdx[j]) != vusX[i].end()) {
-				vXsData[j * rowNum + i] = 1;
-			}
-		}
-	}
-	finish = clock();
-	cout << "Calc xs time: " << (double)(finish - mid) / CLOCKS_PER_SEC << " s" << endl;
+	// for (auto& w : vvWd[260]) cout << w << '\t' << w.size() << endl;
+	//for (auto& w : vDic) cout << w << '\t' << w.size() << endl;
 
+	// vector<double> vSum(colNum);
+	// for (int i = 0; i < rowNum; ++i) {
+	// 	for (int j = 0; j < colNum; ++j) {
+	// 		vSum[j] += vX[j * rowNum + i];
+	// 	}
+	// }
+	// for (auto val : vSum) cout << val << endl;
 	// 测试输出
 	// cout << rowNum << '\t' << colNum << endl;
 	// ofstream ofs1("d:\\result_xsum.txt");
@@ -186,16 +179,14 @@ void mexFunction(int nlhs, mxArray* plhs[], int nrhs, const mxArray* prhs[]) {
 
 	/* 写入结果 */
 	mid = clock();
-	if (nlhs > 0) {
-		plhs[0] = writeToMatDouble(vXSum.data(), 1, vXSum.size());
+	if (nlhs > 0) { // vvX
+		plhs[0] = writeToMatDouble(vX.data(), rowNum, colNum);
 	}
-	if (nlhs > 1) { // xs
-		plhs[1] = writeToMatDouble(vXsData.data(), rowNum, colNum);
-	}
-	finish = clock();
-	cout << "Write result time: " << (double)(finish - mid) / CLOCKS_PER_SEC << " s" << endl;
 
-	cout << "Calc word occurrence in Dic Total time: " << (double)(finish - begin) / CLOCKS_PER_SEC << " s" << endl;
+	finish = clock();
+	if (flagPrint == 2) cout << "Write result time: " << (double)(finish - mid) / CLOCKS_PER_SEC << " s" << endl;
+
+	if (flagPrint) cout << "Calc word occurrence in Dic Total time: " << (double)(finish - begin) / CLOCKS_PER_SEC << " s" << endl;
 }
 
 // 供c++调试用
diff --git a/MexFunc/main.cpp b/MexFunc/main.cpp
index 876c07e..a355ccf 100644
--- a/MexFunc/main.cpp
+++ b/MexFunc/main.cpp
@@ -13,12 +13,12 @@ int main(int argc, const char** argv)
 	const mxArray* prhs[argReserveNum];
 
 	/* SortDedup */
-	int nlhs = 1, nrhs = 2;
-	MATFile* pwdMat = matOpen("D:\\tmp\\wd_small.mat", "r");
-	prhs[0] = matGetVariable(pwdMat, "wd");
-	prhs[1] = mxCreateString("D:\\Twirls\\runtime\\output_1.dat");
-	prhs[2] = mxCreateDoubleMatrix(1, 1, mxREAL);
-	*mxGetPr(prhs[2]) = 2;
+	// int nlhs = 1, nrhs = 2;
+	// MATFile* pwdMat = matOpen("D:\\tmp\\wd_small.mat", "r");
+	// prhs[0] = matGetVariable(pwdMat, "wd");
+	// prhs[1] = mxCreateString("D:\\Twirls\\runtime\\output_1.dat");
+	// prhs[2] = mxCreateDoubleMatrix(1, 1, mxREAL);
+	// *mxGetPr(prhs[2]) = 2;
 
 	/* CalcEntropy */
 	// int nlhs = 2, nrhs = 4;
@@ -32,12 +32,14 @@ int main(int argc, const char** argv)
 	// *mxGetPr(prhs[3]) = 2;
 
 	/* IsWordInDic  */
-	// MATFile* pwdMat, * pdicMat;
-	// int nlhs = 2, nrhs = 2;
-	// pwdMat = matOpen("D:\\tmp\\wd_large.mat", "r");
-	// pdicMat = matOpen("D:\\tmp\\G_dc_large.mat", "r");
-	// prhs[0] = matGetVariable(pwdMat, "wd"); //鑾峰彇.mat鏂囦欢閲岄潰鍚嶄负matrixName鐨勭煩闃�
-	// prhs[1] = matGetVariable(pdicMat, "dc");
+	MATFile* pwdMat, * pdicMat;
+	int nlhs = 1, nrhs = 3;
+	pwdMat = matOpen("D:\\tmp\\ws_small.mat", "r");
+	pdicMat = matOpen("D:\\tmp\\x_small.mat", "r");
+	prhs[0] = matGetVariable(pwdMat, "ws"); //鑾峰彇.mat鏂囦欢閲岄潰鍚嶄负matrixName鐨勭煩闃�
+	prhs[1] = matGetVariable(pdicMat, "x");
+	prhs[2] = mxCreateDoubleMatrix(1, 1, mxREAL);
+	*mxGetPr(prhs[2]) = 2;
 
 	/* ClusterRandSim */
 	// int nlhs = 2, nrhs = 4;