hydeenoble/csv-normalizer

csv-normalizer

How to Build and Run

• Clone the Repo

git clone https://github.com/hydeenoble/csv-normalizer.git

• Install dependencies

npm install

• Run application

node app.js < /path/to/csv/file.csv

Assumptions on the Input file

• Input CSV is UTF-8
• First row contains column headers
• Column names are: Timestamp, Address, ZIP, FullName, FooDuration, BarDuration, TotalDuration, Notes

Problem Statement

Please write a tool that reads a CSV formatted file on stdin and emits a normalized CSV formatted file on stdout. Normalized, in this case, means:

• The entire CSV is in the UTF-8 character set.
• The Timestamp column should be formatted in ISO-8601 format.
• The Timestamp column should be assumed to be in US/Pacific time; please convert it to US/Eastern.
• All ZIP codes should be formatted as 5 digits. If there are less than 5 digits, assume 0 as the prefix.
• All name columns should be converted to uppercase. There will be non-English names.
• The Address column should be passed through as is, except for Unicode validation. Please note there are commas in the Address field; your CSV parsing will need to take that into account. Commas will only be present inside a quoted string.
• The columns FooDuration and BarDuration are in HH:MM:SS.MS format (where MS is milliseconds); please convert them to a floating point seconds format.
• The column "TotalDuration" is filled with garbage data. For each row, please replace the value of TotalDuration with the sum of FooDuration and BarDuration.
• The column "Notes" is free form text input by end-users; please do not perform any transformations on this column. If there are invalid UTF-8 characters, please replace them with the Unicode Replacement Character.

You can assume that the input document is in UTF-8 and that any times that are missing timezone information are in US/Pacific. If a character is invalid, please replace it with the Unicode Replacement Character. If that replacement makes data invalid (for example, because it turns a date field into something unparseable), print a warning to stderr and drop the row from your output.

You can assume that the sample data we provide will contain all date and time format variants you will need to handle.