tf.compat.v1.gfile.Copy  |  TensorFlow v2.16.1 (original) (raw)

tf.compat.v1.gfile.Copy

Stay organized with collections Save and categorize content based on your preferences.

Copies data from src to dst.

tf.compat.v1.gfile.Copy(
    oldpath, newpath, overwrite=False
)

with open("/tmp/x", "w") as f: f.write("asdf") `` 4 tf.io.gfile.exists("/tmp/x") True tf.io.gfile.copy("/tmp/x", "/tmp/y") tf.io.gfile.exists("/tmp/y") True tf.io.gfile.remove("/tmp/y")

You can also specify the URI scheme for selecting a different filesystem:

with open("/tmp/x", "w") as f: f.write("asdf") `` 4 tf.io.gfile.copy("/tmp/x", "file:///tmp/y") tf.io.gfile.exists("/tmp/y") True tf.io.gfile.remove("/tmp/y")

Note that you need to always specify a file name, even if moving into a new directory. This is because some cloud filesystems don't have the concept of a directory.

with open("/tmp/x", "w") as f: f.write("asdf") `` 4 tf.io.gfile.mkdir("/tmp/new_dir") tf.io.gfile.copy("/tmp/x", "/tmp/new_dir/y") tf.io.gfile.exists("/tmp/new_dir/y") True tf.io.gfile.rmtree("/tmp/new_dir")

If you want to prevent errors if the path already exists, you can useoverwrite argument:

with open("/tmp/x", "w") as f: f.write("asdf") `` 4 tf.io.gfile.copy("/tmp/x", "file:///tmp/y") tf.io.gfile.copy("/tmp/x", "file:///tmp/y", overwrite=True) tf.io.gfile.remove("/tmp/y")

Note that the above will still result in an error if you try to overwrite a directory with a file.

Note that you cannot copy a directory, only file arguments are supported.