Sending special characters to devices using Python scripts

[hfakoor222]

I'm writing some scripts, and i noticed when sending special characters, namely the dash sign "-", to show commands on Cisco devices I am getting bad outputs.

for example:

sh_data_corr = "show data-corruption"

try:
corruption = ConnectHandler.send_command(sh_data_corr)

the output is:

% Invalid input detected at '^' marker.

This is due to the dash sign. I forget how the commands are sent over the wire to the cisco device. Was it in hex? How should i convert my above command to make it work then?

Thanks

[Flavio Miranda · Accepted solution]

Hello,

 Did you check if you are on the right prompt inside the Cisco device?

 This error could happen if you were in 'config t'  even if the command were right.

[hfakoor222]

All my commands are show commands, the rest seem to work fine, except the

| include   -    symbol 

commands below don't return any output, and 

show data-corruption     returns the error at mark output.

sh_bfd = "show bfd neighbors"
sh_bootvar = "sh bootvar | include BOOT variable does not exist|CONFIG_FILE does not exist"
# shows if we have boot system image & boot config file saved in memory caches
sh_cap_route = "sh capability feature routing | section Disabled"
# outputs disabled routing features
sh_clock = "show clock" # default cisco clock is 24Hour"
sh_cpu_report = "sh monitor event-trace cpu-report brief back 110:00"
# needs to be configured with process cpu on node or return empty string
sh_data_corr = "show data-corruption" # checks data inconsistency errors
sh_logging = "show logging | include [A-Z]+-[0-4]" #shows log severity 3 or above
#we will extract logging level 0-4 in the last 5 days, let the node parse the regex
sh_mem_failure = "sh memory failures alloc | include \S+.*[0-9]"
#the [0-9] returns any memory failures detected, or an empty string
sh_int_down = "sh ip int brief" # regex 'if err-disabled'
sh_cef = "show cef state"
sh_mem_total = "show memory free totals | include (Processor).*[0-9]" # parses second line of memory cache

I changed PyCharms global and file encodings to both ASCII and UT-8 and it doesn't seem to solve it.

[hfakoor222]

also sh_mem_total   seems to work as the dash symbol is inside brackets:  [-]

but 

sh_logging = "show logging | include [A-Z]+-[0-4]

returns error at mark output  as dash symbol is outside of brackets. I'm thinking it  might be parsing the - symbol as a range?

[bigevilbeard]

I think i saw this or something like it years back with Napalm. In the end i created an alias command on the device. As you noted the other stuff works, and i found work arounds for the pipe command. Dirty trick, but it did work.

[hfakoor222]

That's really interesting. For now I want to avoid aliases. I might try plain Python IDE instead of Pycharm if I can't find a way for it to work.

I tried it in the Python console from the system, and same exact issue. The best guess I have on my mind right now is default encodings in windows not doing what i want it to do. Not being able to use special characters takes a lot of usability of of the scripts, so I have to find a workaround

[hfakoor222]

After all that work and yes, I should've been under enable. it is working now.