Tensorflow使用sparse_column_with_hash_bucket()和sparse_column_with_keys()对特征进行处理，爆出Table未初始化的问题：

在初始化OP中加入tf.tables_initializer()后解决：

init_op = [tf.global_variables_initializer(), tf.local_variables_initializer(), 
           tf.tables_initializer()]
with tf.Session() as sess:
    sess.run(init_op)

tf.tables_initializer：Returns an Op that initializes all tables of the default graph.