add:diff_csv

diff_csv.py

@@ -0,0 +1,47 @@
+import pandas as pd
+
+
+def read_largest_column(file_path: str) -> int:
+    csv_file = file_path
+    largest_column_count = 0
+    with open(csv_file, 'r') as temp_f:
+        lines = temp_f.readlines()
+        for l in lines:
+            column_count = len(l.split(',')) + 1
+            # 找到列数最多的行
+            largest_column_count = column_count if largest_column_count < column_count else largest_column_count
+    temp_f.close()
+    # colunm_names为最大列数展开
+
+    return largest_column_count
+
+
+def cut_csv_to_model(file_path: str, pd_names=None):
+    df = pd.read_csv(file_path, names=pd_names)
+    # 获取文件总行数
+    row_num = len(df)
+    # 确定每个小文件要包含的数据量
+    step = 1000
+    for start in range(0, row_num, step):
+        stop = start + step
+        filename = f"cut/{file_path.split('.')[0]}_{start}-{stop}.csv"
+        d = df[start: stop]
+        print("Saving file : " + filename + ", data size : " + str(len(d)))
+        d.to_csv(filename, index=None)
+
+
+file_path = "CICNTTor_browsing_raw.csv"
+# file_path = "cut/CICNTTor_browsing_raw_0-1000.csv"
+column_count = read_largest_column(file_path)
+column_names = [i for i in range(0, column_count)]
+
+# cut_csv_to_model(file_path, pd_names=column_names)
+
+df = pd.read_csv(file_path, header=None, delimiter=',', names=column_names)
+df_l = df[df.columns[:8]]
+df_r = df[df.columns[8:]]
+df_r_l1 = df_r.shift(periods=-1,axis=1)
+df_result = df_r_l1 - df_r
+df_result = pd.concat([df_l,df_result],axis=1)
+# df_result =
+df_result.to_csv('CICNTTor_browsing_diff.csv', index=False)