From f2c16080715c20aff19cba9f9c64c389ff875a20 Mon Sep 17 00:00:00 2001
From: Jeff Larson <thejefflarson@gmail.com>
Date: Mon, 9 Jan 2012 14:43:58 -0500
Subject: [PATCH 1/2] implement a loosey option for cases in which the schema
 doesn't need to be strict

---
 csvkit/sql.py              | 19 ++++++++++---------
 csvkit/utilities/csvsql.py | 11 ++++++++---
 2 files changed, 18 insertions(+), 12 deletions(-)

diff --git a/csvkit/sql.py b/csvkit/sql.py
index 4203d54f4..0d89c9f3d 100644
--- a/csvkit/sql.py
+++ b/csvkit/sql.py
@@ -22,7 +22,7 @@
 
 NULL_COLUMN_MAX_LENGTH = 32
 
-def make_column(column):
+def make_column(column, loosey=False):
     """
     Creates a sqlalchemy column from a csvkit Column.
     """
@@ -45,12 +45,13 @@ def make_column(column):
     else:
         raise ValueError('Unexpected normalized column type: %s' % column.type)
 
-    if column.type is NoneType:
-        sql_type_kwargs['length'] = NULL_COLUMN_MAX_LENGTH
-    elif column.type is unicode:
-        sql_type_kwargs['length'] = column.max_length()
+    if loosey is False:
+        if column.type is NoneType:
+            sql_type_kwargs['length'] = NULL_COLUMN_MAX_LENGTH
+        elif column.type is unicode:
+            sql_type_kwargs['length'] = column.max_length()
 
-    sql_column_kwargs['nullable'] = column.has_nulls()
+        sql_column_kwargs['nullable'] = column.has_nulls()
 
     return Column(column.name, sql_column_type(**sql_type_kwargs), **sql_column_kwargs)
 
@@ -60,7 +61,7 @@ def get_connection(connection_string):
 
     return engine, metadata
 
-def make_table(csv_table, name='table_name', metadata=None):
+def make_table(csv_table, name='table_name', loosey=False, metadata=None):
     """
     Creates a sqlalchemy table from a csvkit Table.
     """
@@ -70,7 +71,7 @@ def make_table(csv_table, name='table_name', metadata=None):
     sql_table = Table(csv_table.name, metadata)
 
     for column in csv_table:
-        sql_table.append_column(make_column(column))
+        sql_table.append_column(make_column(column, loosey))
 
     return sql_table
 
@@ -82,7 +83,7 @@ def make_create_table_statement(sql_table, dialect=None):
         module = __import__('sqlalchemy.dialects.%s' % DIALECTS[dialect], fromlist=['dialect'])
         sql_dialect = module.dialect()
     else:
-        sql_dialect = None 
+        sql_dialect = None
 
     return unicode(CreateTable(sql_table).compile(dialect=sql_dialect)).strip() + ';'
 
diff --git a/csvkit/utilities/csvsql.py b/csvkit/utilities/csvsql.py
index 801a9a369..e2ba41079 100644
--- a/csvkit/utilities/csvsql.py
+++ b/csvkit/utilities/csvsql.py
@@ -22,6 +22,8 @@ def add_arguments(self):
                             help='In addition to creating the table, also insert the data into the table. Only valid when --db is specified.')
         self.argparser.add_argument('--table', dest='table_name',
                             help='Specify a name for the table to be created. If omitted, the filename (minus extension) will be used.')
+        self.argparser.add_argument('--loosey', dest='loosey', action='store_true',
+                            help='Generate a schema without limits or null checks. Useful for big tables.')
 
     def main(self):
         if self.args.table_name:
@@ -32,6 +34,9 @@ def main(self):
         else:
             self.argparser.error('The --table argument is required when providing data over STDIN.')
 
+        if self.args.loosey:
+            loosey = True
+
         if self.args.dialect and self.args.connection_string:
             self.argparser.error('The --dialect option is only valid when --db is not specified.')
 
@@ -47,7 +52,7 @@ def main(self):
             except ImportError:
                 raise ImportError('You don\'t appear to have the necessary database backend installed for connection string you\'re trying to use.. Available backends include:\n\nPostgresql:\tpip install psycopg2\nMySQL:\t\tpip install MySQL-python\n\nFor details on connection strings and other backends, please see the SQLAlchemy documentation on dialects at: \n\nhttp://www.sqlalchemy.org/docs/dialects/\n\n')
 
-            sql_table = sql.make_table(csv_table, table_name, metadata)
+            sql_table = sql.make_table(csv_table, table_name, loosey, metadata)
             sql_table.create()
 
             if self.args.insert:
@@ -55,11 +60,11 @@ def main(self):
                 headers = csv_table.headers()
 
                 for row in csv_table.to_rows(serialize_dates=True):
-                    engine.execute(insert, [dict(zip(headers, row)), ]) 
+                    engine.execute(insert, [dict(zip(headers, row)), ])
 
         # Writing to file
         else:
-            sql_table = sql.make_table(csv_table, table_name)
+            sql_table = sql.make_table(csv_table, table_name, loosey)
             self.output_file.write((u'%s\n' % sql.make_create_table_statement(sql_table, dialect=self.args.dialect)).encode('utf-8'))
 
 if __name__ == '__main__':

From 628ffbe03f6a99997290701a262685a36e4dd532 Mon Sep 17 00:00:00 2001
From: Jeff Larson <thejefflarson@gmail.com>
Date: Mon, 9 Jan 2012 14:46:46 -0500
Subject: [PATCH 2/2] docs for --loosey option

---
 docs/scripts/csvsql.rst | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/docs/scripts/csvsql.rst b/docs/scripts/csvsql.rst
index 07443757a..290675c9a 100644
--- a/docs/scripts/csvsql.rst
+++ b/docs/scripts/csvsql.rst
@@ -39,6 +39,8 @@ Generate SQL statements for a CSV file or create execute those statements direct
                             specified.
       --table TABLE_NAME    Specify a name for the table to be created. If
                             omitted, the filename (minus extension) will be used.
+      --loosey              Generate a schema without limits or null checks.
+                            Useful for big tables.
 
 Also see: :doc:`common_arguments`.