1、直接执行.sql脚本


import numpy as np import pandas as pd import lightgbm as lgb from pandas import DataFrame from sklearn.model_selection import train_test_split from io import StringIO import gc import sys import os hive_cmd = "hive -f ./sql/sql.sql" output = os.popen(hive_cmd) data_cart_prop = pd.read_csv(StringIO(unicode(output.read(),'utf-8')), sep="\t",header=0)



2、Hive语句执行

假如有如下hive sql：

hive_cmd = 'hive -e "select count(*) from hbase.routermac_sort_10;"'

一般在python中按照如下方式执行该hive sql：

os.system(hive_cmd)

---------------------



hive_cmd1 = "hive -f ./user.sql" output1 = os.popen(hive_cmd1) test_user = pd.read_csv(StringIO(unicode(output1.read(),'utf-8')), sep="\t",header=0)  hive_cmd2 = "hive -f ./action.sql" output2 = os.popen(hive_cmd2) test_action = pd.read_csv(StringIO(unicode(output2.read(),'utf-8')), sep="\t",header=0)  hive_cmd3 = "hive -f ./click.sql" output3 = os.popen(hive_cmd3) test_click = pd.read_csv(StringIO(unicode(output3.read(),'utf-8')), sep="\t",header=0)


为了显示表头，在脚本中加上一句：set hive.cli.print.header=true;

或者，使用如下语句：


hive_cmd = 'hive -e "set hive.cli.print.header=true;SELECT * FROM dev.temp_dev_jypt_decor_user_label_phase_one_view_feature WHERE(dt = "2018-09-17");"' output = os.popen(hive_cmd) data_cart_prop = pd.read_csv(StringIO(unicode(output.read(),'utf-8')), sep="\t",header=0)





3、tf 显存占用


import tensorflow as tf tf.enable_eager_execution() x = tf.get_variable('x', shape=[1], initializer=tf.constant_initializer(3.)) with tf.GradientTape() as tape:            y = tf.square(x)     y_grad = tape.gradient(y, x)         print([y.numpy(), y_grad.numpy()])