TypeError: coercing to Unicode: need string or buffer, file found

Question

I'm getting a buffer error, but I have a line buffer in the 'with open()' block at line 123. Is this the correct location for the buffer? Should I have something in the connects class too/instead, or the parser argument maybe. I'm trying to use a file with addresses or hostnames using a '--host_file' argument to run the SNMP commands on multiple ASA's. It works using the '--host' argument just fine. Any help provided would be greatly appreceated. ``` Traceback (most recent call last): File "asaos-snmpv3-tool.py", line 145, in <module> main() File "asaos-snmpv3-tool.py", line 124, in main with open(hosts, mode='r', buffering=-1): TypeError: coercing to Unicode: need string or buffer, file found import pexpect import argparse PROMPT = ['# ', '>>> ', '>', '\$ '] SNMPGROUPCMD = ' snmp-server group ' V3PRIVCMD = ' v3 priv ' SNMPSRVUSRCMD = ' snmp-server user ' V3AUTHCMD = ' v3 auth ' PRIVCMD = ' priv ' SNMPSRVHOSTCMD = ' snmp-server host ' VERSION3CMD = ' version 3 ' SHAHMACCMD = ' sha ' S…

Accepted Answer

There are a whole series of problems here:

```
if hosts:
    with open(hosts, mode='r', buffering=1):
```

Here, `hosts` is clearly meant to be a filename. You open that file… and then don't store it in a variable anywhere, and go on to use the filename as if it were a file.

But, before you even get there, `hosts` is not actually a filename. One of the nifty features of `argparse` is that it can automatically open files for you, and you're asking it to do that by using `type=file` in the argument spec. Which means that `hosts` is actually a file object, and you're trying to use that file object itself as a filename! Which is what' actually causing the `TypeError: coercing to Unicode: need string or buffer, file found`: `open` tries to convert filenames to Unicode strings, and it has no idea how to do that with a file object. If you want to `open` hosts, don't tell argparse to do it for you; just leave its type as a `str`.

```
        for line in hosts:
```

Since `hosts` is supposed to be a filename, as a string, this will give you each character in the filename, as a string. Which is not useful.

```
            hosts = line.rstrip
```

This overwrites the `hosts` variable, which you need. And, even worse, it overwrites it with the bound method `line.rstrip`, rather than the result of calling `line.rstrip`.

```
            child = connects(user, hosts, passwd, en_passwd)
```

And here, you're passing that bound method to a function that wants a string, which isn't going to work.

Don't keep using the same variable name over and over to mean different things. Use a different name for each thing. Ideally one whose name relates in some way to the thing it's holding. For example, a ech (stripped) line in a file named `hosts` is probably a `host`, or a `hosts_entry`… one thing it's not is all of your `hosts`.

Anyway, to fix all of these problems:

```
if hosts:
    for line in hosts:
        host = line.rstrip()
        child = connects(user, host, passwd, en_passwd)
```

Or, if you want to control how `hosts` is being opened (I'm not sure why you think it's important to specify `buffering=1`, but presumably you have some reason?):

```
parser.add_argument('--host_file', dest='hosts', type=str, help='specify a target host file')

# …

if hosts:
    with open(hosts, buffering=1) as hosts_file:
        for line in hosts_file:
            host = line.rstrip()
            child = connects(user, host, passwd, en_passwd)
```

TypeError: coercing to Unicode: need string or buffer, file found

Solution

Problem