results 0.1.1552206909

Don't get mad, get results

Don't get mad, get results

Tabular data and SQL for people who don't have time to faff about.

Move between xlsx, xls, csv, python, postgres and back with ease.

Features:

• Zero-boilerplate database creating, connecting and querying.
• Loading/tidying/transforming csv and excel data.
• Autodetect column types, load your data with little or no manual specification.
• Powerful multi-column, multi-order keyset paging of database results.
• Schema syncing.

Limitations

• Python 3.6+, PostgreSQL 10+ only. Many features will work with other databases, but many won't. Just use Postgres!

Installation

results is on PyPI. Install it with pip or any of the (many) Python package managers.

Scenario

Somebody gives you a messy csv or excel file. You need to load it, clean it up, put it into a database, query it, make a pivot table from it, then send the pivot table to somebody as a csv.

results is here to get this sort of thing done quickly and with minimum possible fuss.

Let's see.

First, load and clean:

import results

# load a csv (in this example, some airport data)
sheet = results.from_file("airports.csv")

# do general cleanup
sheet.standardize_spaces()
sheet.set_blanks_to_none()

# give the keys lowercase-with-underscore names to keep the database happy
cleaned = sheet.with_standardized_keys()

Then, create a database:

# create a database
DB = "postgresql:///example"

db = results.db(DB)

# create it if it doesn't exist
db.create_database()

Then create a table for the data, automatically guessing the columns and creating a table to match.

# guess the column types
guessed = cleaned.guessed_sql_column_types()

# create a table for the data
create_table_statement = results.create_table_statement("data", guessed)

# create or auto-update the table structure in the database
# syncing requires a copy of postgres running locally with your current user set up as superuser
db.sync_db_structure_to_definition(create_table_statement, confirm=False)

Then insert the data and freely query it.

# insert the data. you can also do upserts with upsert_on!
db.insert("data", cleaned)

# show recent airfreight numbers from the top 5 airports
# ss means "single statement"
query_result = db.ss(
    """
with top5 as (
    select
        foreignport, sum(freight_in_tonnes)
    from
        data
    where year >= 2010
    group by
        foreignport
    order by 2 desc
    limit 5
)

select
    year, foreignport, sum(freight_in_tonnes)
from
    data
where
    year >= 2010
    and foreignport in (select foreignport from top5)
group by 1, 2
order by 1, 2

"""
)

Create a pivot table, then print it as markdown or save it as csv.

# create a pivot table
pivot = query_result.pivoted()

# print the pivot table in markdown format
print(pivot.md)

Output:

|   year |   Auckland |    Dubai |   Hong Kong |   Kuala Lumpur |   Singapore |
|-------:|-----------:|---------:|------------:|---------------:|------------:|
|   2010 |     288997 | 145527   |      404735 |       226787   |      529407 |
|   2011 |     304628 | 169868   |      428990 |       244053   |      583921 |
|   2012 |     312828 | 259444   |      400596 |       272093   |      614155 |
|   2013 |     306783 | 257263   |      353895 |       272804   |      592886 |
|   2014 |     309318 | 244776   |      330521 |       261438   |      620419 |
|   2015 |     286202 | 263378   |      290292 |       252906   |      633862 |
|   2016 |     285973 | 236419   |      309556 |       175858   |      614172 |
|   2017 |     314405 | 226048   |      340216 |       199868   |      662505 |
|   2018 |     126712 |  91611.2 |      134540 |        74667.5 |      250653 |

Save the table as a csv:

pivot.save_csv("2010s_freight_sources_top5.csv")

Design philosophy

• Avoid boilerplate at all costs. Make it as simple as possible but no simpler.

• Don't reinvent the wheel: results uses sqlalchemy for database connections, existing excel parsing libraries for excel parsing, etc etc. results brings it all together, sprinkles some sugar on top, and puts it at your fingertips.

• Eat your own dogfood: We use this ourselves every day.

Documentation

This README.md is currently all there is :( But we'll add more soon, we promise!

Credits

• Rob
• Jason
• Nick

Contributions

Yes please!