csv-etl
A rules based approach to performing ETL operations on csv files.
Installation
Requires Python 3
Using pip
python -m venv venv
source venv/bin/activate
pip install csv-etl
Locally
git clone https://github.com/winslowdibona/csv-etl.git
cd csv-etl
python -m venv venv
source venv/bin/activate
pip install --editable .
Developing
git clone https://github.com/winslowdibona/csv-etl.git
cd csv-etl
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
Overview
The goal of this project is to provide a re-usable way to perform ETL operations on csv files. This implementation takes a given row of data from a csv file and applies a set of rules to generate a new format of the data.
Rule Overview
The Rule
in this project has the following properties
target
source
type
input_type
output_type
operations
Using different combinations of values for these properties can give us a flexible toolset for extracting data.
str
target - Once we have the extracted the value from the csv and applied our rule to it, the resulting value will appear under this value as the key in the result
str/list
source - This is where we want to pull the data from in the csv. If the source is a string, it will fetch that single column. If the source is a list, it will fetch all of the values.
RuleType
type - This can be one of two options
Static
Calculation
If RuleType.Static, the rule will simply return the value stored in Rule.source under Rule.target
If RuleType.Calculation, the rule will fetch the value(s) defined in Rule.source, and perform the operations on them
InputType
input_type - The data type you would like to read the value in from the csv as.
This can be one of three options
String
Integer
Decimal
OutputType
output_type - The data type you would like the resulting value to be.
This can be one of four options
String
Integer
Decimal
Date
list
operations - A list of strings that will be run through python eval
statement. The value(s) extracted from the csv will be available for use in these operations. If there is only a single source, the value will be assigned to the variable s
. If there are multiple sources, the values will be passed in as a list under the variable s
.
Defining Rules
These rules can be defined programmatically, or via a YAML configuration. The configuraiton follows the below structure
rules:
-
target: target_name
type: Static || Calculation
input_type: String || Integer || Decimal
output_type: String || Integer || Decimal || Date
source: source_name || [source, names]
operations: ['operations', 'to', 'run']
Converting CSV Data
Once we have a set of rules, we can use the CSVConverter
class to execute our rules on a data set.
from csv_etl import CSVConverter
csv_converter = CSVConverter(rules)
result = csv_converter.convert('path/to/csv_file')
This will give us back a list of dictionaries, with each item in the list representing the modified data for each row in the initial csv file.
We can also get the result fed back to us in csv format as a string
result = csv_converter.convert('path/to/csv/file', to='csv')
Usage
CLI
csv-etl ./examples/order_data/config.yaml ./examples/order_data/test_data.csv
$ csv-etl --help
Usage: csv-etl [OPTIONS] CONFIG CSV
Options:
--outfile TEXT File path to write the result to
--format TEXT Format the result should be. "json" or "csv"
--help Show this message and exit.
Examples
More detailed usage and programmatic examples are provided
Helpful Make Commands
Generating Documentation
make docs
open html/csv_etl/index.html
Running Tests
make pytest
Getting Test Coverage
make test-cov
open cov_html/index.html
What Next?
Better error handling
Currently the errors are just printed to the console. Might be nice to have an option to gather them and have them represented in the resulting data set somehow.
Better handling of eval statements
Right now the use of eval
statements is a little tailored for the original problem
and could use some more exploration on how they can be further utilized and how to handle
potential errors.
Also right now the ability to perform multiple operations on a single extracted value is doable, but performing multiple operations on multiple extracted values will not. eval
statements don't allow assignment of variables, and this prevents us from performing some logic, and being able to use the resulting values again in a list. Further exploration here may find a way to do multiple operations with multiple extracted values.