如何从日期时间字符串中检索年，月，日，小时和分钟？

Question

I am trying to extract from a tensor string "2018/12/31 22:59" the values of year, month, day, hour and minute. I found this function tf.string_split for this task, but my code is throwing an error

Traceback (most recent call last):
  File "path/to/my/file.py", line 12, in <module>
    date = split_date_time[0]
TypeError: 'SparseTensor' object does not support indexing

Here's the code

import tensorflow as tf

date_time = tf.placeholder(dtype=tf.string)

day = tf.placeholder(shape=[None], dtype=tf.int32),
month = tf.placeholder(shape=[None], dtype=tf.int32),
year = tf.placeholder(shape=[None], dtype=tf.int32),
hour = tf.placeholder(shape=[None], dtype=tf.int32),
minute = tf.placeholder(shape=[None], dtype=tf.int32)

split_date_time = tf.string_split(date_time, ' ')
date = split_date_time[0]
time = split_date_time[1]

date_splitted = tf.string_split(date, '-')
year = date_splitted[0]
month = date_splitted[1]
day = date_splitted[2]

time_spplitted = tf.string_split(time, ':')
hour = time_spplitted[0]
minute = time_spplitted[1]

with tf.Session() as sess:
    print (sess.run(year, feed_dict={date_time: "2018-12-31 22:59"}))
    print (sess.run(month, feed_dict={date_time: "2018-12-31 22:59"}))
    print (sess.run(day, feed_dict={date_time: "2018-12-31 22:59"}))
    print (sess.run(hour, feed_dict={date_time: "2018-12-31 22:59"}))
    print (sess.run(minute, feed_dict={date_time: "2018-12-31 22:59"}))

Answer 1

You have several problems in your code (mainly because you apparently didn't read any documentation regarding the functions you're trying to use). I will just mention a few crucial problems related to the specific problem you're trying to solve (but I strongly encourage you to study the basics of TensorFlow and its computational model).

First, as the documentation of tf.string_split states, the first argument of tf.string_split should be a " 1-D string Tensor, the strings to split ". However, "2018-12-31 22:59" is a 0-D string tensor.

Second, tf.string_split returns a tf.SparseTensor , which cannot be indexed!

Here's a possible solution to your problem:

import tensorflow as tf

date_time = tf.placeholder(shape=(1,), dtype=tf.string)

split_date_time = tf.string_split(date_time, ' ')

date = split_date_time.values[0]
time = split_date_time.values[1]

split_date = tf.string_split([date], '-')
split_time = tf.string_split([time], ':')

year = split_date.values[0]
month = split_date.values[1]
day = split_date.values[2]

hours = split_time.values[0]
minutes = split_time.values[1]

with tf.Session() as sess:
    year, month, day, hours, minutes = sess.run([year, month, day, hours, minutes],
                                                feed_dict={date_time: ["2018-12-31 22:59"]})
    print("Year =", year)
    print("Month =", month)
    print("Day =", day)
    print("Hours =", hours)
    print("Minutes =", minutes)