Merge branch 'master' of http://git.wanmeizhensuo.com/ML/ffm-baseline

tensnsorflow/multi.py

@@ -5,13 +5,32 @@ from pyspark.conf import SparkConf
 import pytispark.pytispark as pti
 from pyspark.sql import SQLContext
 from pyspark.sql import SparkSession
+import datetime
 
 
 def test():
     spark = SparkSession.builder.enableHiveSupport().getOrCreate()
     ti = pti.TiContext(spark)
     ti.tidbMapDatabase("jerry_test")
-    spark.sql("select device_id from esmm_pre_data limit 6").show(6)
+    df = spark.sql("select max(stat_date) from esmm_train_data")
+    df.show()
+    t = df.rdd.map(lambda x: str(x[0])).collect()
+    print(t.count())
+
+
+
+    # data = [(0, 18.0), (1, 19.0), (2, 8.0), (3, 5.0), (4, 2.2), (5, 9.2), (6, 14.4)]
+    # df = spark.createDataFrame(data, ["id", "hour"])
+    # df.show(6)
+    # t = df.rdd.map(lambda x:x[0]).collect()
+    # print(t)
+
+    # validate_date = spark.sql("select max(stat_date) from esmm_train_data").rdd.map(lambda x: str(x[0]))
+    # print(validate_date.count())
+    # print("validate_date:" + validate_date)
+    # temp = datetime.datetime.strptime(validate_date, "%Y-%m-%d")
+    # start = (temp - datetime.timedelta(days=10)).strftime("%Y-%m-%d")
+    # print(start)
 
 
 if __name__ == '__main__':