csv-source-of-truth

Collection of modules and documentation to enable using Microsoft Excel (and CSV files) as a Source of Truth for configuration data.

published

Overview

The goal of this effort is to enable using a on-line or local spreadsheet program (Microsoft Excel) to define the configuration of a network fabric. The use case which prompted this development is automating the configuration of Cisco Application Centric Infrastructure (ACI) using Ansible. However, these modules have value to other devices in the network infrastructure.

Why use a Spreadsheet?

Microsoft Excel is readily available, provides a high degree of functionality for data visualization and analysis, and is commonly used by network engineers for the definition of data center fabrics.

While YAML is a human-readable data serialization language and perhaps more suitable, especially given using Ansible as a configuration management tool, the whitespace indentation to provide structure can be confusing initially to the non-programmer.

Data Structure

Spreadsheets represent data in a tabular data format similar to a relational database. The database stores data in objects called tables, which consist of columns and rows. The columns include a column name and other attributes.

The spreadsheet file (.xlsx) is analogous to a database, while one or more sheets in the file are analogous to tables in the database.

In our use case, configuring a Cisco ACI fabric, the Cisco Application Policy Infrastructure Controller (Cisco APIC) manages the model of the ACI fabric in a hierarchical (tree) structure. At the top of the tree structure is the root (topRoot) and the policy Universe (polUni) is contained by topRoot.

The challenge in using a spreadsheet to represent the configuration of the ACI fabric focuses on optimizing and eliminating redundancy in the tabular format of the sheet to a hierarchal structure of the APIC Management Information Tree (MIT).

Elimination of Manual Operations

While the network administrator can manually issue a File -> Save As in Excel to create the CSV (“comma-separated values”) files needed for the work-flow, this manual operation can be done programmatically.

The benefits of developing a programmatic interface include the following:

  • Sheet names are verified to be valid file names
  • Column headers verified to be valid variable names
  • The playbook specifies and extracts only the necessary sheet(s) from the spreadsheet file
  • Eliminates human error in saving sheets

The spreadsheet file can be stored in a SCM (Source Code Management System) with the Ansible playbooks, or downloaded by the playbook using the uri module.

What is Included?

This repository contains the following modules in the /library/ directory:

  • xls_to_csv.py Reads an Excel .xlsx file and output .csv files for each sheet specified
  • csv_to_facts.py Reads a CSV file and returns as ansible facts a list of dictionaries for each row

Also, a filter plugin in /plugins/filter directory:

  • to_csv.py Called with a list of dictionaries and outputs a string (with newlines) which can be written to a CSV file

The two modules can be executed in sequence. For example, xls_to_csv.py can be used to extract sheets from a spreadsheet into individual CSV files, and in a subsequent task or play, csv_to_facts.py can be used to expose the data from a CSV file as variables to a playbook. Either module can be used independently of the other.

The module csv_to_facts.py was originally developed in 2015. It remains at the original (now deprecated) location. The design goal of these modules subscribes to the UNIX philosophy of

Make each program do one thing well. To do a new job, build afresh rather than complicate old programs by adding new “features”.

which is why xls_to_csv.py was developed as a separate module.

Installation

If your intent is to use these modules with Ansible Tower or with a multi-user system, you may which to install them in a common location on an existing system. You can also use an ephemeral environment using Vagrant.

Requirements

Python packages pandas and xlrd are required if using library/xls_to_csv.py. They can be installed using Pip.

$ sudo pip install xlrd pandas

or if using Python3

$  sudo pip3 install xlrd pandas

How to Test the Software

If you have Ansible Tower installed on an existing system you can download the modules to a shared directory and modify the Ansible configuration file to identify the location.

Alternately, you can create an ephemeral test environment using Vagrant.

Install to an Existing System

Refer to the instructions for Adding modules and plugins locally.

The modules library/xls_to_csv.py and library/csv_to_facts.py can be written to the Ansible ‘magic’ directories. Modify the /etc/ansible/ansible.cfg file to include:

library        = /usr/share/ansible/

Then use wget to download these modules to that directory:

$ cd /usr/share/ansible
$ sudo wget https://raw.githubusercontent.com/joelwking/csv-source-of-truth/master/library/csv_to_facts.py
$ sudo wget https://raw.githubusercontent.com/joelwking/csv-source-of-truth/master/library/xls_to_csv.py
$ chmod 755 *.py

To verify installation, issue ansible-doc csv_to_facts.

The filter plugin can be written to the Ansible ‘magic’ directories as well. Refer to the documentation on developing locally.

Modify the ansible.cfg file to include the location where the filter plugin to_csv.py is downloaded. For example:

filter_plugins     = /usr/share/ansible/plugins/filter 

Install using Vagrant

Alternately, if you only wish to create a test environment using Vagrant, there is a sample configuration file in files/vagrant/. After issuing the vagrant up command using the Vagrantfile in this repository, vagrant ssh and complete the configuration with the following commands:

$ git clone https://github.com/joelwking/csv-source-of-truth.git
$ export ANSIBLE_LIBRARY=$HOME/csv-source-of-truth/library
$ export ANSIBLE_DEPRECATION_WARNINGS=False
$ sudo pip install xlrd pandas
$ sudo pip3 install xlrd pandas
$ cd csv-source-of-truth

Note: This README file was tested using the Ansible 2.9.13 release.

Usage

This section illustrates using the modules to extract and manipulate data used for configuring a Cisco ACI fabric. A sample spreadsheet is available in files/aci/. You may wish familiarize yourself with the contents of that spreadsheet prior to completing the following examples.

Note: The inventory file ./files/inventory/yml defines a group name of python, which has two localhosts, one for Python 2.7.17 and Python 3.6.9. You do not need to run the sample playbooks using both versions, it is to verify the code functions using either version.

Create CSV

The module library/xls_to_csv.py reads an Excel .xlsx file and writes .csv file(s).

The sample input file files/aci/ACI_DHCP_configuration.xlsx contains two sheets, “DHCP Relay”, and “data_centers”.

The module is executed as an Ansible ad-hoc command to verify the sheet names. Run the module with an empty string for the sheets argument and the warn option to verify the sheet names. The module reports the names of the sheets located in the source file. By using the host name of python and the inventory.yml file, both Python 2 and Python 3 are verified.

[email protected]:~/csv-source-of-truth$ ansible python -m xls_to_csv -a "src='$HOME/csv-source-of-truth/files/aci/ACI_DHCP_configuration.xlsx' dest=/tmp sheets='' warn=true" -i ./files/inventory.yml
/usr/lib/python2.7/dist-packages/ansible/parsing/vault/__init__.py:44: CryptographyDeprecationWarning: Python 2 is no longer supported by the Python core team. Support for it is now deprecated in cryptography, and will be removed in a future release.
  from cryptography.exceptions import InvalidSignature
[WARNING]: sheet "DHCP Relay" found in source file, skipping
[WARNING]: sheet "data_centers" found in source file, skipping
python2 | SUCCESS => {
    "ansible_facts": {
        "sheet_filenames": []
    },
    "changed": false
}
python3 | SUCCESS => {
    "ansible_facts": {
        "sheet_filenames": []
    },
    "changed": false
}

Because no sheets were written, the changed flag is set to ‘false’.

Naming Requirement

The module converts the sheet names and column headers to valid file and variable names. Ansible variable names must be letters, numbers and underscores and must start with a letter. The module removes special characters (other than an underscore) and spaces.

Run the module as an ad-hoc command and specifying the sheet “DHCP Relay”. Note the module removed the embedded space in the sheet named “DHCP Relay” and exported the contents to /tmp/DHCPRelay.

[email protected]:~/csv-source-of-truth$ ansible python -m xls_to_csv -a "src='$HOME/csv-source-of-truth/files/aci/ACI_DHCP_configuration.xlsx' dest=/tmp sheets='DHCP Relay'" -i ./files/inventory.yml
/usr/lib/python2.7/dist-packages/ansible/parsing/vault/__init__.py:44: CryptographyDeprecationWarning: Python 2 is no longer supported by the Python core team. Support for it is now deprecated in cryptography, and will be removed in a future release.
  from cryptography.exceptions import InvalidSignature
python2 | CHANGED => {
    "ansible_facts": {
        "sheet_filenames": [
            "DHCPRelay"
        ]
    },
    "changed": true
}
python3 | CHANGED => {
    "ansible_facts": {
        "sheet_filenames": [
            "DHCPRelay"
        ]
    },
    "changed": true
}

Individual sheets in the spreadsheet file are extracted and written to file(s). The module returns a variable sheet_filenames which is a list filenames of the extracted sheets saved in the destination directory.

Executing from a Playbook

This section illustrates running the xls_to_csv module from a playbook.

Review test_xls.yml. The xls_to_csv module is executed specifying the source spreadsheet file, a destination directory to write the selected sheets, and a list of sheets to extract from the spreadsheet file.

- hosts: python
  gather_facts: no 
  connection: local
  tags: [play1, xls_to_csv]

  vars:
    spreadsheet: '{{ playbook_dir }}/files/aci/ACI_DHCP_configuration.xlsx'  
    dest: '{{ playbook_dir }}/files/aci/'

  tasks:
    - name: Extract the sheets from the Excel file, creating CSV files 
      xls_to_csv:
        src: '{{ spreadsheet }}' 
        dest: '{{ dest }}' 
        sheets: 
          - 'DHCP Relay'
          - 'data_centers'
        warn: True

    - debug:
        msg: 'File {{ dest }}{{ item }}.csv has been created'
      loop: '{{ sheet_filenames }}'
      tags: [debug]

Tip: As a best practice, specify your source and target directories relative to the playbook directory, for example: ‘{{ playbook_dir }}/files’ as the destination directory.

Running the module as an Ansible ad-hoc command, we can identify the names of the sheets in the spreadsheet file. The sheets variable is a list of sheets we wish to extract from the spreadsheet file and write the result to individual CSV files. Execute the test_xls.yml playbook and specify the tag play1 which identifies the first play in the YAML file.

This step extracts the selected sheets and writes them to the destination directory.

[email protected]:~/csv-source-of-truth$ ansible-playbook ./test_xls.yml --tags play1 -i ./files/inventory.yml
/usr/lib/python2.7/dist-packages/ansible/parsing/vault/__init__.py:44: CryptographyDeprecationWarning: Python 2 is no longer supported by the Python core team. Support for it is now deprecated in cryptography, and will be removed in a future release.
  from cryptography.exceptions import InvalidSignature

PLAY [python] *****************************************************************************************************************

TASK [Extract the sheets from the Excel file, creating CSV files] **************************************************************
changed: [python2]
changed: [python3]

TASK [debug] *******************************************************************************************************************
ok: [python3] => (item=DHCPRelay) => {
    "msg": "File /home/vagrant/csv-source-of-truth/files/aci/DHCPRelay.csv has been created"
}
ok: [python3] => (item=data_centers) => {
    "msg": "File /home/vagrant/csv-source-of-truth/files/aci/data_centers.csv has been created"
}
ok: [python2] => (item=data_centers) => {
    "msg": "File /home/vagrant/csv-source-of-truth/files/aci/data_centers.csv has been created"
}
ok: [python2] => (item=DHCPRelay) => {
    "msg": "File /home/vagrant/csv-source-of-truth/files/aci/DHCPRelay.csv has been created"
}

...

PLAY RECAP *************************************************************************************************************************
python2                    : ok=2    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0
python3                    : ok=2    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

Next review the contents of the output file(s).

Determine the Column Header Names

The column headers of the CSV file(s) can be identified by looking at the first record in each output file.

[email protected]:~/csv-source-of-truth$ head -1 ./files/aci/*.csv
==> ./files/aci/DHCPRelay.csv <==
DC,Tenant,VRF,BD,AppProfile,EPG,VLAN,DHCP,L3Out

==> ./files/aci/data_centers.csv <==
DC,Tenant,fullname,address1,city,state,postalcode
[email protected]:~/csv-source-of-truth$

Because the column headers are used as variable names by playbooks, issue the head -1 command for each file to identify the column headers.


Tip: Knowing the column headers is important for converting the CSV file to Ansible facts in the csv_to_facts module described later.

Summary

This section illustrates programmatically extracting one or more sheets from a spreadsheet file. This enables the network engineer to use a spreadsheet program, like Microsoft Excel, to define the initial state of the network infrastructure. The xls_to_csv module is used to identify and extract the sheet names in the spreadsheet file, writing the contents of the sheet to a CSV formatted file.

By reviewing the column headers in the output files, the cell contents of each row can be used as configuration data for the network infrastructure.

Read CSV

The module library/csv_to_facts.py reads a CSV file and returns as Ansible facts a list of dictionaries for each row. The column header is the key, the contents of the cell is the value.

For example, assume there are two data centers, (DC1 and DC2) each with an APIC Clusters managing the policy of their respective domain. Each data center will have one or more tenants, each tenant may have one or more VRFs, Bridge Domains, and so on. The tabular data, therefore, will have redundant information at the root of the tree.

Looking at the sample data (file DHCPRelay.csv), there are 15 rows of data, representing two data centers, with each data center having two tenants.

While there are a number of columns in each row that describe the configuration of the ACI fabric(s), for this explanation we will focus on the column DC (data center), and Tenant.

[email protected]:~/csv-source-of-truth$ cat ./files/aci/DHCPRelay.csv | cut -d ',' -f 1,2
DC,Tenant
DC1,XXV-INT
DC1,XXV-DMZ
DC2,XXV-DMZ
DC1,XXV-INT
DC1,XXV-INT
DC1,XXV-INT
DC1,XXV-INT
DC1,XXV-INT
DC1,XXV-INT
DC1,XXV-INT
DC1,XXV-INT
DC2,XXV-INT
DC2,XXV-INT
DC2,XXV-INT
DC2,XXV-INT

Execute csv_to_facts

The module csv_to_facts reads a CSV file and returns as Ansible facts a list of dictionaries for each row. The column header is the key, the contents of the cell is the value.

The default behavior of csv_to_facts is to return the content of the source (src) file in a list variable named spreadsheet. Examine this playbook sample.

- hosts: python
  gather_facts: no
  tags: [play2, csv_to_facts]

  tasks:
    - name: Default behavior of csv_to_facts
      csv_to_facts:
        src: '{{ playbook_dir }}/files/aci/DHCPRelay.csv'
    - debug:
        msg: '{{ item.DC }} {{ item.Tenant }}'
      loop: '{{ spreadsheet }}'

By executing the playbook, we iterate over the list variable spreadsheet and reference the DC and Tenant columns. The list variable spreadsheet has a length of 15, each item corresponds to a row in the CSV file.

Execute play2 to verify the debug module can reference the variables specified in the CSV file.

[email protected]:~/csv-source-of-truth$ ansible-playbook ./test_xls.yml --tags play2 -i ./files/inventory.yml
/usr/lib/python2.7/dist-packages/ansible/parsing/vault/__init__.py:44: CryptographyDeprecationWarning: Python 2 is no longer supported by the Python core team. Support for it is now deprecated in cryptography, and will be removed in a future release.
  from cryptography.exceptions import InvalidSignature

PLAY [python] ******************************************************************************************************************

PLAY [python] ******************************************************************************************************************

TASK [Default behavior of csv_to_facts] ****************************************************************************************
ok: [python2]
ok: [python3]

TASK [debug] ***********************************************************************************************************************
ok: [python3] => (item={u'BD': u'BD-BOX-RAZOR', u'AppProfile': u'AP-PRD', u'VLAN': u'39', u'DC': u'DC1', u'L3Out': u'ERN-N7KCORESW-L3OUT', u'VRF': u'VRF-XXV-INT', u'DHCP': u'Yes', u'EPG': u'EPG-BOX-RAZOR', u'Tenant': u'XXV-INT'}) => {
    "msg": "DC1 XXV-INT"
}

Note: The output of the above execution was truncated for brevity, only the first row was shown.

Tip: The argument table can be specified to provide a value other than the default value of spreadsheet. Use ansible-doc csv_to_facts for more details.

Optimizing Tabular Data

Because a spreadsheet represents data in a tabular format, and the ACI fabric configuration is stored in a Management Information Tree (a hierarchical) structure, there will be repetitive data defined in the sheet.

The APIC REST API itself is idempotent, issuing a request to create (or delete) a tenant with the same name does not add a duplicate tenant, it simply validates the tenant name exists (or in the case of a delete, it doesn’t exist), and returns a successful status (200 OK) to the caller.

However, this is inefficient, causing unnecessary API calls and increasing the run time and memory usage of the playbook.

If the data were loaded into a relational database, SQL provides a SELECT DISTINCT statement to return only distinct (different) values. Relational database tables often contain duplicate values. In this example the goal is to return only the unique (distinct) values.

In SQL, the statement might look like the following:

SELECT DISTINCT DC, Tenant FROM DHCPRelay;

In the next section, csv_to_facts can be configured to return distinct values for the columns specified.

Create virtual spreadsheets

The csv_to_facts module can optimize the input file by creating a variable which eliminates redundancy by returning unique (distinct) values for the columns specified.

This optimization of the input CSV file is a accomplished by creating multiple views of the input data with the specified columns and returning a list of unique values for the column specified.

This is accomplished by manipulating the input spreadsheet and using the Python set data type to create unordered collections of unique elements.

These virtual spreadsheets can be used to loop (iterate) over a task. Ansible uses loops (or with_items) to repeat a task multiple times.

As an example, assume there is a task file (tenant in role ansible-aci-tenant) which creates an ACI tenant, the goal is to iterate over this task, creating (or deleting) the Tenants for data center DC1 specified in the CSV file.

  vars:
    data_center: 'DC1'
    change_request: 'CHG00012345'

  tasks:
    - name: Manage Tenants
      include_role:
        name: ansible-aci-tenant
        tasks_from: tenant
      vars:
        fvTenant:
          name: '{{ item.Tenant }}'
          descr: '{{ change_request }}'
      loop: '{{ TENANTs }}'
      when:
        - item.DC == data_center

To accomplish this task efficiently, our list variable TENANTs should be made up of the unique tenants in each data center. Each data center has the same two tenants, XXV-INT and XXV-DMZ.

The goal is to create a list variable TENANTs with only four entries from the 15 rows in the original spreadsheet.

{
    "ansible_facts": {
        "TENANTs": [
            {
                "DC": "DC2",
                "Tenant": "XXV-INT"
            },
            {
                "DC": "DC1",
                "Tenant": "XXV-DMZ"
            },
            {
                "DC": "DC2",
                "Tenant": "XXV-DMZ"
            },
            {
                "DC": "DC1",
                "Tenant": "XXV-INT"
            }
        ]
}

By using the vsheets argument, specify a list of virtual spreadsheets you wish to create, and for each virtual spreadsheet, specify a list of columns. The aforementioned list of TENANTs can be created using the following example:

- hosts: localhost
  gather_facts: no
  tags: [play3, csv_to_facts]

  tasks:
    - name: Create virtual spreadsheet of data centers (DC) and tenants
      csv_to_facts:
        src: '{{ playbook_dir }}/files/aci/DHCPRelay.csv'
        vsheets: 
          - TENANTs:
             - DC
             - Tenant

Executing the play with verbose mode (-v) illustrates the variable TENANTs is a list of unique values of columns DC and Tenant.

[email protected]:~/csv-source-of-truth$ ansible-playbook ./test_xls.yml --tags play3 -i ./files/inventory.yml -v
/usr/lib/python2.7/dist-packages/ansible/parsing/vault/__init__.py:44: CryptographyDeprecationWarning: Python 2 is no longer supported by the Python core team. Support for it is now deprecated in cryptography, and will be removed in a future release.
  from cryptography.exceptions import InvalidSignature
Using /etc/ansible/ansible.cfg as config file

PLAY [python] *****************************************************************************************************************

PLAY [python] *****************************************************************************************************************

PLAY [python] *****************************************************************************************************************

TASK [Create virtual spreadsheet of data centers (DC) and tenants] *************************************************************
ok: [python2] => {"ansible_facts": {"TENANTs": [{"DC": "DC2", "Tenant": "XXV-INT"}, {"DC": "DC1", "Tenant": "XXV-DMZ"}, {"DC": "DC2", "Tenant": "XXV-DMZ"}, {"DC": "DC1", "Tenant": "XXV-INT"}], "spreadsheet": [{"AppProfile": "AP-PRD", "BD": "BD-BOX-RAZOR", "DC": "DC1", "DHCP": "Yes", "EPG": "EPG-BOX-RAZOR", "L3Out": "ERN-N7KCORESW-L3OUT", "Tenant": "XXV-INT", "VLAN": "39", "VRF": "VRF-XXV-INT"}

Note: The output of the above execution was truncated for brevity.

Sheet Filenames

The module xls_to_csv returns the variable sheet_filenames. This variable can be referenced in subsequent tasks or plays to identify the names of the files created.

By executing plays 1 and 4, the variable sheet_filenames can be referenced in subsequent play(s). Execute these plays by specifying their tags.

[email protected]:~/csv-source-of-truth$  ansible-playbook ./test_xls.yml --tags 'play1, play4' -i ./files/inventory.yml
/usr/lib/python2.7/dist-packages/ansible/parsing/vault/__init__.py:44: CryptographyDeprecationWarning: Python 2 is no longer supported by the Python core team. Support for it is now deprecated in cryptography, and will be removed in a future release.
  from cryptography.exceptions import InvalidSignature

PLAY [python] ******************************************************************************************************************

TASK [Extract the sheets from the Excel file, creating CSV files] **************************************************************
changed: [python2]
changed: [python3]

TASK [debug] *******************************************************************************************************************
ok: [python3] => (item=DHCPRelay) => {
    "msg": "File /home/vagrant/csv-source-of-truth/files/aci/DHCPRelay.csv has been created"
}
ok: [python3] => (item=data_centers) => {
    "msg": "File /home/vagrant/csv-source-of-truth/files/aci/data_centers.csv has been created"
}
ok: [python2] => (item=data_centers) => {
    "msg": "File /home/vagrant/csv-source-of-truth/files/aci/data_centers.csv has been created"
}
ok: [python2] => (item=DHCPRelay) => {
    "msg": "File /home/vagrant/csv-source-of-truth/files/aci/DHCPRelay.csv has been created"
}

PLAY [python] ******************************************************************************************************************

PLAY [python] ******************************************************************************************************************

PLAY [python] ******************************************************************************************************************

TASK [Create summarized virtual sheets, loading the variables in a namespace using the filename] *******************************
ok: [python2] => (item=data_centers)
ok: [python3] => (item=DHCPRelay)
ok: [python2] => (item=DHCPRelay)
ok: [python3] => (item=data_centers)

TASK [debug] *******************************************************************************************************************
ok: [python3] => (item={u'address1': u'6007 Applegate Lane', u'DC': u'DC2', u'Tenant': u'XXV-DMZ'}) => {
    "msg": {
        "DC": "DC2",
        "Tenant": "XXV-DMZ",
        "address1": "6007 Applegate Lane"
    }
}

Note: The output of the above execution was truncated for brevity.

Sample Use Case

To illustrate the use of both the xls_to_csv and the csv_to_facts modules,

Review the playbook manage_aci_dhcp.yml. This playbook contains two plays.

The first play uses xls_to_csv reading the input file located in files/aci/ACI_DHCP_configuration.xlsx, selects sheet ‘DHCP Relay’ and writes it to a CSV file in the same directory with a file name of ‘DHCPRelay.csv’

The second play uses csv_to_facts. This module creates a list variable named DHCPentries which includes these columns from the sheet ‘DHCP Relay’.

    - DC
    - Tenant
    - BD
    - AppProfile
    - DHCP
    - EPG

The second task (Associate multiple DHCP servers with a Tenant, Bridge Domain) in the play iterates over the list variable DHCPentries and all elements of dhcp.server, which is defined in files/aci_dhcp.yml.

Extra vars are passed from the command line to select the appropriate data center and test if the variable DHCP is True (specified as ‘Yes’ in the sheet).

In this example, the debug module is used as a placeholder for the appropriate Ansible ACI module(s). For more information on using Ansible to configure an ACI fabric, please refer to the Cisco ACI Guide.

[email protected]:~/csv-source-of-truth$ ansible-playbook -i ./files/inventory.yml ./manage_aci_dhcp.yml -e "data_ce
> nter=DC1 ticket=CHG58234"
/usr/lib/python2.7/dist-packages/ansible/parsing/vault/__init__.py:44: CryptographyDeprecationWarning: Python 2 is no longer supported by the Python core team. Support for it is now deprecated in cryptography, and will be removed in a future release.
  from cryptography.exceptions import InvalidSignature

PLAY [localhost] ***************************************************************************************************************

TASK [Extract the sheets from the Excel file, creating CSV file(s)] ************************************************************
[WARNING]: sheet "data_centers" found in source file, skipping
changed: [localhost]

PLAY [APIC] ********************************************************************************************************************

TASK [Summarize the sheet and include as facts] ********************************************************************************
ok: [aci-demo.sandbox.wwtatc.local]

TASK [Associate multiple DHCP servers with a Tenant, Bridge Domain] ************************************************************
ok: [aci-demo.sandbox.wwtatc.local] => (item=[{u'BD': u'BD-BOX-PRVLIN1', u'AppProfile': u'AP-BOX', u'DHCP': u'Yes', u'EPG': u'EPG-BOX-PRVLIN1', u'DC': u'DC1', u'Tenant': u'XXV-INT'}, {u'value': {u'dn': u'uni/tn-WWT_NULL/ap-MANAGEMENT/epg-DHCP', u'addr': u'198.51.100.17'}, u'key': u'DHCP-DC2-PRD'}]) => {
    "msg": "Apply to tenant=XXV-INT BD=BD-BOX-PRVLIN1 DHCP Label=DHCP-DC2-PRD server=198.51.100.17 dn=uni/tn-WWT_NULL/ap-MANAGEMENT/epg-DHCP"
}
ok: [aci-demo.sandbox.wwtatc.local] => (item=[{u'BD': u'BD-BOX-PRVLIN1', u'AppProfile': u'AP-BOX', u'DHCP': u'Yes', u'EPG': u'EPG-BOX-PRVLIN1', u'DC': u'DC1', u'Tenant': u'XXV-INT'}, {u'value': {u'dn': u'uni/tn-common/out-L3-ATC/instP-L3-ATC', u'addr': u'203.0.113.17'}, u'key': u'DHCP-DC1-PRD'}]) => {
    "msg": "Apply to tenant=XXV-INT BD=BD-BOX-PRVLIN1 DHCP Label=DHCP-DC1-PRD server=203.0.113.17 dn=uni/tn-common/out-L3-ATC/instP-L3-ATC"
}
skipping: [aci-demo.sandbox.wwtatc.local] => (item=[{u'BD': u'BD-BOX-VMNFS2', u'AppProfile': u'AP-BOX', u'DHCP': u'Yes', u'EPG': u'EPG-BOX-VMNFS2', u'DC': u'DC2', u'Tenant': u'XXV-INT'}, {u'value': {u'dn': u'uni/tn-WWT_NULL/ap-MANAGEMENT/epg-DHCP', u'addr': u'198.51.100.17'}, u'key': u'DHCP-DC2-PRD'}])

Note: The output of the above execution was truncated for brevity.

Tip: Between play 1 and play 2, the variable sheet is a different format, the embedded space has been eliminated by module xls_to_csv. This illustrates the importance of creating spreadsheet and column names that do not contain spaces or special characters.

For example, because “extra vars” take precedence, specifying sheet on the command line will cause the playbook to fail. To illustrate execute the following command:

[email protected]:~/csv-source-of-truth$ ansible-playbook -i ./files/inventory.yml ./manage_aci_dhcp.yml -e "data_center=DC1 sheet='DHCP Relay' ticket=CHG58234"

Summary

This section illustrates reading one or more CSV files into variables. Additionally, module csv_to_facts can be configured to return virtual sheets which return only distinct (different) values. This feature implements the behavior of the SQL command DISTINCT, eliminating redundant rows and optimizing playbook execution.

Generate CSV

The filter plugin to_csv.py expects a list of dictionaries as input. Each list item is converted to a row (line) of the CSV file
The dictionary keys are inserted before the first record, representing the column headers. Output is a string with newline characters. This string can be written to a file and opened with Excel or other program which recognizes CSV file formats.

#!/usr/bin/env ansible-playbook
---
- name: create CSV file
  hosts: localhost
  connection: local
  gather_facts: no

  vars:
    list_of_dict:
      - {'name': 'Bob', 'title': 'programmer'}
      - {'name': 'Sally', 'title': 'writer'}
      - {'name': 'Russell', 'title': 'thought leader'}

  tasks:
    - name: Write the output to a file in CSV format
      copy:
        content: '{{ list_of_dict | to_csv }}'
        dest: '/tmp/query_api.csv'
        mode: 0644

The resulting output file /tmp/query_api.csv looks as follows:

name,title
Bob,programmer
Sally,writer
Russell,thought leader

Summary

Using the filter plugin to_csv.py aids in creating CSV files. Many network engineers are familiar with using the CSV format to define and visualize the network infrastructure.

Synopsis

Organizing and managing the Source of Truth for configuring infrastructure is equally important to developing the process (workflow) to effect the configuration. Network engineers commonly use spreadsheets to represent network infrastructure which organize data in a tabular data format similar to a relational database.

The modules and playbooks shown in this repository can be used to manipulate and import tabular data for managing infrastructure devices.

Known Issues

None.

Getting Help

If you have questions or comments on network and infrastructure automation in general or these modules, please contact the author or [email protected].

Credits and References

Author

Joel W. King (@joelwking) [email protected] – Engineering and Innovations, Network Solutions, World Wide Technology

GitHub

View Github