Convert CSV to SQL and create a table in your Drupal installation

CSV are a great way to give a client to enter their data that will be migrated to the site. However, for the migration itself, SQL would be quicker.

Install the drush command by running drush dl csv2sql.

Execute the command: drush csv2sql /PATH/TO/file.csv

You have the option to scan a whole directory instead of giving an exact path to a csv file. Execute the command: drush csv2sql /PATH/TO/directory/

Will create a _raw_file table in the Drupal installation which drush is running under.

• Each column is created as varchar 255 by default. However it is possible to override it by setting the header in the CSV file.
• To prevent the user of MySql reserved words, each column is prefixed with an underscore (e.g. _order)
• The first column is treated as the primary column
• In the SQL a serial __id column is created
• Index can be added to any column by adding index:TRUE to the column's header, The first column of each table is added to the index by default unless stated otherwise (index:FALSE) in the header, This will make for a faster migration if you need to use any column as key for referencing other entities.

The complex column will be translated in the DB to an amount column type int(11) where NULL value is allowed.

The values that can be passed in the header are the ones that are expected by db_create_table()

In order to deploy a local copy to a remote server you may need to export the SQL tables, and later import them. Here's a handy bash command to do it - it exports all the tables prefixed with _raw_ into the raws.sql file. The second comamnd simply imports that SQL file to the remote server.

# Export SQL tables into a file.
drush sqlq "SHOW TABLES LIKE '_raw_%';" | awk -v ORS=, '{ print $1 }' | sed 's/,$//' | sed 's/^Tables_in_[^,]*,//' | drush sql-dump > raws.sql

# Replace the drush alias with your own.
`drush @remote-alias sql-connect` < raws.sql