My dict appears to overwrite and not update with new rows?

Question

Given the following code:

from ciscoconfparse import CiscoConfParse
from pprint import pprint
parse = CiscoConfParse("testconfig.txt", syntax="junos")

interfaces = {}
intfs = parse.find_objects_w_parents(r'^interface', r'^\s+ge-')
for intfobj in intfs:
    intf_name = intfobj.text.strip()
    interfaces.update({'name': intf_name})
    descr = intfobj.re_match_iter_typed(r'description\s+"(\S.+?)"$', group=1)
    interfaces.update({'description': descr})
    mode = intfobj.re_match_iter_typed(r'port-mode\s+(\S+)\s*$', group=1,
        all_children=True)
    interfaces.update({'mode': mode})


print (interfaces)

Using the following test data:

https://bpaste.net/show/df422a96aaae

interfaces {
    ge-2/0/0 {
        description "site1;;hostname1;ge-16/0/9;;;TRUST;";
        unit 0 {
            family ethernet-switching {
                port-mode trunk;
            }
        }
    }
    ge-2/0/2 {
        description "site2;;hostname2;ge-16/0/8;;;TRUST;";
        unit 0 {
            family ethernet-switching {
                port-mode trunk;
            }
        }
    }

    vstp {
        bpdu-block-on-edge;
        vlan VLAN_0005 {
            interface ge-2/0/0 {
                edge;
            }
        }
        vlan VLAN_0015 {
            interface ge-2/0/0 {
                edge;
            }
            interface ge-2/0/2 {
                edge;                   
            }
        }  
    }
}

I'm trying to understand why my interfaces = {} just contains one row:

{'name': 'ge-2/0/2', 'description': 'site2;;hostname2;ge-16/0/8;;;TRUST;', 'mode': 'trunk'}

I would expect it to contain both interfaces from the test data:

{'name': 'ge-2/0/0', 'description': 'site1;;hostname1;ge-16/0/9;;;TRUST;', 'mode': 'trunk'}
{'name': 'ge-2/0/2', 'description': 'site2;;hostname2;ge-16/0/8;;;TRUST;', 'mode': 'trunk'}

`interfaces` is a dict. It can't contain multiple "rows". You need a list of dicts or similar. — Michael Butscher, Sep 24 '18 at 22:59
`dict`s are mappings, not sequences. Each key is associated (or mapped) to a single value. That "single" value can be a `list` or other nested container, however. — martineau, Sep 24 '18 at 23:08
Any chance you have an example? I'm without a doubt getting confused about things. Interfaces is the dict, within the dict each ge-x/x/x should be a new key with values of desciption and port-mode — ddevalco, Sep 24 '18 at 23:11

score 1 · Accepted Answer · answered Sep 24 '18 at 23:14

You are creating a dict and updating its values. So you end up with a dict with last values. You could rather use a list of dicts, as such :

from ciscoconfparse import CiscoConfParse
from pprint import pprint
parse = CiscoConfParse("testconfig.txt", syntax="junos")

interfaces = []
intfs = parse.find_objects_w_parents(r'^interface', r'^\s+ge-')
for intfobj in intfs:
    interface = {}
    interface['name'] = intfobj.text.strip()
    interface['description'] = intfobj.re_match_iter_typed(r'description\s+"(\S.+?)"$', group=1)
    interface['mode'] = intfobj.re_match_iter_typed(r'port-mode\s+(\S+)\s*$', group=1,
        all_children=True)
    interfaces.append(interface)

print (interfaces)

score 0 · Answer 2 · answered Sep 24 '18 at 23:19

Could try something like this

allInterfaces = {}
interfaces = {}
intfs = parse.find_objects_w_parents(r'^interface', r'^\s+ge-')
for intfobj in intfs:
    intf_name = intfobj.text.strip()
    interfaces.update({'name': intf_name})
    descr = intfobj.re_match_iter_typed(r'description\s+"(\S.+?)"$', group=1)
    interfaces.update({'description': descr})
    mode = intfobj.re_match_iter_typed(r'port-mode\s+(\S+)\s*$', group=1,
        all_children=True)
    interfaces.update({'mode': mode})
    allInterfaces = allInterfaces + [interfaces,]

print (allInterfaces)

Ooops... not fast enough :-) – Chai Ang Sep 24 '18 at 23:20 — Chai Ang, Sep 24 '18 at 23:20

My dict appears to overwrite and not update with new rows?

2 Answers2