change test file

tensnsorflow/multi.py

@@ -128,18 +128,19 @@ def con_sql(db,sql):
     return df
 
 def get_filename(dir_in):
-    pre_path = "hdfs://172.16.32.4:8020"
-    args = "hdfs dfs -ls " + dir_in + " | awk '{print $8}'"
-    proc = subprocess.Popen(args, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=True)
-    s_output, s_err = proc.communicate()
-    all_dart_dirs = s_output.split()
-    a = []
-    for i in all_dart_dirs:
-        b = str(i).split("/")[4]
-        if b[:4] == "part":
-            tmp = pre_path + str(i)[2:-1]
-            a.append(tmp)
-    return a
+    pre_add = "hdfs://172.16.32.4:8020/strategy/esmm/"
+    x = []
+    for i in range(0,200):
+        if i < 10:
+            t = pre_add+dir_in+"/part-r-0000"+str(i)
+            x.append(t)
+        elif 10 <= x < 100:
+            t = pre_add + dir_in + "/part-r-000" + str(i)
+            x.append(t)
+        elif 100 <= x < 200:
+            t = pre_add + dir_in + "/part-r-00" + str(i)
+            x.append(t)
+    return x
 
 if __name__ == '__main__':
     # sparkConf = SparkConf().set("spark.hive.mapred.supports.subdirectories", "true") \
@@ -170,9 +171,9 @@ if __name__ == '__main__':
     # df.show()
     # df.createOrReplaceTempView("df")
     # t = spark.sql("select id from df").map()
-    tr_files = get_filename("/strategy/esmm/tr")
+    tr_files = get_filename("tr")
     print(tr_files)
-    va_files = get_filename("/strategy/esmm/va")
+    va_files = get_filename("va")
     print("test")
     print(va_files)