Skip to content

Latest commit

 

History

History
46 lines (32 loc) · 1.73 KB

formatting_tips.md

File metadata and controls

46 lines (32 loc) · 1.73 KB
Raw

Tips for formatting/cleaning CSV

These are tips for cleaning a CSV that, coupled with a corresponding TMCF, represents graph data suitable for import into Data Commons.

Options to use when writing CSV

When you open the CSV file for writing, it is convenient to escape the double-quote character occurring within a field. You can do that by using the following args to CSV writer (example).

doublequote=False, escapechar='\\'

Emitting DC string values

DC strings use double quotes, as does the default CSV quoting. So CSV fields that store DC String values need to be escaped.

For example, to represent name: "Foo Bar", the CSV field value should be "\"Foo Bar\"".

You can achieve that by opening the output CSV to auto-escape quotes (as noted above), and then in python code use double-quotes without escaping, like '"Foo Bar"'.

Emitting DC repeated values

DC repeated values use comma as the delimiter, as does the default CSV field delimiter. So CSV fields that store repeated DC values should be quoted.

For example, to represent typeOf: dcs:County, dcs:State in MCF, the cell value will have "dcs:County, dcs:State".

Just use the CSV writer to write the field value, and it will introduce the double quotes.

You can have combinations of repeated strings with some containing commas. For example, to represent name: "Google, Inc.", "Alphabet, Inc.", the CSV field value would be "\"Google, Inc.\", \"Alphabet, Inc.\"".

You can achieve that by opening the output CSV to auto-escape quotes (as noted above), and then in python code use double-quotes without escaping, like '"Google. Inc.", "Alphabet, Inc."'.