csv2table

A fast and flexible command line tool to automate parsing and importing of CSV files into database tables. Currently the only database supported is MySQL.

Use case

Every night you have to fetch a set of CSV files from a remote SFTP server and import each file into a corresponding mysql table. If the destination table doesn't exist, automatically create it.

There may be several columns that require special processing:

• expire_date: normal values are sent as dd.mm.yyyy, but there is a special value that signifies that there's no upper limit in time, 31.12.2099. This value must be saved as a database NULL value
• expire_date: besides processing the value, an index is needed on this column for faster searching
• amount: convert from a specific locale (e.g. 12,5) to the database locale (e.g. 12.5)
• start_date: a more complex example, this column can have 2 types of values: 1) dd.mm.yyyy specifying an exact date or 2) an integer n meaning the number of months from January 1st of current year. In the later case we must calculate the the exact date and save it in database. (NOTE: Such complex example will be possible only after integrating an embedded language. See "Planned features" section).

Documentation

Table and column names transformations

Table and column names are sanitized and transformed before creating the corresponding database table. The following rules apply:

• white space is trimmed
• inner white space and - are replaced with _
• special characters that can be part of the CSV header but can't be part of a database name are removed: <>/\()
• "umlaut" characters are replaced by their normal counterparts (e.g. ä or á is replaced by a)
• and finally all names are lower cased.

Example: A CSV column named Date of Receipt will be saved in database as date_of_receipt.

Configuration files

All configuration is manually defined in toml files. There are 2 types of configuration files:

1) Global configuration file csv2table.toml

If there is a file named csv2table.toml in the working directory (the directory where csv files are found) it is loaded as the first configuration file. You can skip using the global configuration file, its common usage is when you have many CSV files that share identical configuration options (e.g. the database credentials).

2) CSV specifific configuration file

They have the same name as the csv file but with ".toml" extension. The main purpose of the CSV configuration file is to define the column mapping. Besides column mapping you can also define global configuration options in which case they will overwrite the configuration options defined in the global file.

ATTENTION: The file specific configuration file is mandatory, otherwise the CSV file will not be imported.

Configuration options

Main configuration options:

Option	Description	Default value
host	mysql host name
port	mysql port	3306
db	database name
username	database username
password	database password
table	table name	defaults to "sanitized" name of the CSV file
drop	drop and recreate table before import (true/false)	false
truncate	truncate table before import (true/false)	false
autoPk	create an auto increment PK (int)	false
defaultColType	default column definition	VARCHAR(255) NULL DEFAULT NULL
tableOptions	table options when creating the table	COLLATE='utf8_general_ci' ENGINE=InnoDB
bulkInsertSize	how many rows to insert at once	10000
verbose	verbosity to console	false
email	a section where email notifications cand be configured, see "Email notifications" section

Column mapping

Column mapping is defined in the CSV specific configuration file. Mapping options for each column are grouped under a mapping.column_name table. If we take the example from the "Use case" section, mapping options for column expire_date will be grouped under mapping.expire_date table.

Note: the name of the column is the database column name (see "Table and column names transformations" section).

Column mapping options:

Mapping option	Description	Example	Default value
type	column type	type = "INT NULL DEFAULT NULL"	defaults to global option defaultColType
index	add column index (true/false)	index = true	false
nullIf	set column to DB null if its value is one in the list	nullIf = ["31.12.2999", ""]
nullIfEmpty	set column to DB null if its value is empty string	nullIfEmpty = true	false

format mapping option is used to format a value from the CSV format to DB format. Possible patterns:

Usage	Description	Example
decimal point	hint decimal point by simply assigning a number containing the CSV decimal point	format = "1,2" (hint that "," is the decimal point)
date parsing	parse a date using "Go" language date and time pattern matching	format = "02.01.2006" (date format is dd.mm.yyyy)
time parsing	parse a date/time using "Go" language date and time pattern matching	format = "02.01.2006 15:04:05" (date format is dd.mm.yyyy hh:mm:ss)

Example of mapping options for column expire_date:

[mapping.expire_date]
    type = "DATE NULL DEFAULT NULL"
    format = "02.01.2006"
    index = true

Email notifications

It's possible to enable email notifications through SMTP protocol. Example sending notifications when an error occurs, usig GMail SMTP.

[email]
    sendOnSuccess = false # don't send email if everything ok
    sendOnError = true # send if an error occurs
    from = "me@gmail.com"
    to = ["me@gmail.com"]
    smtpServer = "smtp.gmail.com:587"
[email.plainAuth]
    identity = ""
    username = "me@gmail.com"
    password = "gmail_secret"
    host = "smtp.gmail.com"

TODO documentation: It's possible to overwrite the default emails subject and body as as configuration options.

Full example

sample_import.csv file to be imported

No ID;Reading;Reading_Date;Channel
1;2,5;02.05.2014;X
2;2,5;02.05.2014;X
3;2,5;02.05.2014;First
4;2,5;02.05.2014;X
5;2,5;02.05.2014;Last
6;2,5;07.05.2014;X
7;2,5;17.07.2014;X
8;2,5;23.07.2014;X
9;2,5;28.07.2014;X
10;2,5;17.03.2015;X
11;2,5;01.02.2016;

csv2table.toml global configuration

host = "localhost"
port = 3306
username = "my_username"
password = "my_pwd"
db = "my_db"

# config defaults
verbose = true
drop = true
truncate = true
autoPk = true
bulkInsertSize = 15000

# email 
[email]
    sendOnSuccess = false
    sendOnError = true # send email notification on error
    from = "myself@gmail.com"
    to = ["myself@gmail.com"]
    smtpServer = "smtp.gmail.com:587"
[email.plainAuth]
    identity = ""
    username = "myself@gmail.com"
    password = "gmail_secret"
    host = "smtp.gmail.com"

sample_import.toml file configuration

[mapping.no_id]
    type = "INT NULL DEFAULT NULL"
    index = true
[mapping.reading]
    type = "DOUBLE NULL DEFAULT NULL"
    format = "1,2" 
[mapping.reading_date]
    type = "DATE NULL DEFAULT NULL"
    format = "02.01.2006"
    nullIf = ["31.12.2999"]
[mapping.channel]
    nullIF = ["Last"]
    nullIfEmpty = true

And result in database

Planned features

1. valueIf operator. Usage valueIf = [0, "abc", "def"]
2. Integrate an embedded language to allow complex culomn parsing. Candidates: Tengo, Lua.

License

This project is licensed under the terms of the MIT license.