469,575 Members | 1,680 Online
Bytes | Developer Community
New Post

Home Posts Topics Members FAQ

Post your question to a community of 469,575 developers. It's quick & easy.

variable expansion with sqlite

Hi I'm using SQlite and the CSV module and trying to create a class
that converts data from CSV file into a SQLite table.

My script curently uses functions for everything and I'm trying to
improve my class programming. The problem I'm having is with variable
expansion.

self.cursor.executemany('INSERT INTO test VALUES (?)', CSVinput)

If CSVinput is a tuple with only 1 value, everything is fine. If I
want to use a tuple with more than 1 value, I need to add more
question marks. As I'm writing a class I don't want to hard code a
specific number of ?s into the INSERT statement.

The two solutions I can think of are;
using python subsitution to create a number of question marks, but
this seems very dirty
or
finding someway to substitue tuples or lists into the statement - I'm
not sure if this should be done using Python or SQLite substitution
though.

Any tips on where to start looking?

Thanks, Marc.
Jul 30 '08 #1
4 2190
marc wyburn wrote:
Hi I'm using SQlite and the CSV module and trying to create a class
that converts data from CSV file into a SQLite table.

My script curently uses functions for everything and I'm trying to
improve my class programming. The problem I'm having is with variable
expansion.

self.cursor.executemany('INSERT INTO test VALUES (?)', CSVinput)

If CSVinput is a tuple with only 1 value, everything is fine. If I
want to use a tuple with more than 1 value, I need to add more
question marks. As I'm writing a class I don't want to hard code a
specific number of ?s into the INSERT statement.

The two solutions I can think of are;
using python subsitution to create a number of question marks, but
this seems very dirty
or
finding someway to substitue tuples or lists into the statement - I'm
not sure if this should be done using Python or SQLite substitution
though.

I do this kind of thing sometimes:

<test.csv>
a,b,c
1,2,3
4,5,6
</test.csv>

<code>
import csv
import sqlite3

reader = csv.reader (open ("test.csv", "rb"))
csv_colnames = reader.next ()

db = sqlite3.connect (":memory:")
coldefs = ", ".join ("%s VARCHAR (200)" % c for c in csv_colnames)
db.execute ("CREATE TABLE test (%s)" % coldefs)

insert_cols = ", ".join (csv_colnames)
insert_qmarks = ", ".join ("?" for _ in csv_colnames)
insert_sql = "INSERT INTO test (%s) VALUES (%s)" % (insert_cols, insert_qmarks)

db.executemany (insert_sql, list (reader))
for row in db.execute ("SELECT * FROM test"):
print row

</code>

Obviously, this is a proof-of-concept code. I'm (ab)using
the convenience functions at database level, I'm hardcoding
the column definitions, and I'm making a few other assumptions,
but I think it serves as an illustration.

Of course, you're only a few steps away from something
like sqlalchemy, but sometimes rolling your own is good.

TJG
Jul 30 '08 #2
Tim Golden wrote:
marc wyburn wrote:
>Hi I'm using SQlite and the CSV module and trying to create a class
that converts data from CSV file into a SQLite table.

My script curently uses functions for everything and I'm trying to
improve my class programming. The problem I'm having is with variable
expansion.

self.cursor.executemany('INSERT INTO test VALUES (?)', CSVinput)

If CSVinput is a tuple with only 1 value, everything is fine. If I
want to use a tuple with more than 1 value, I need to add more
question marks. As I'm writing a class I don't want to hard code a
specific number of ?s into the INSERT statement.

The two solutions I can think of are;
using python subsitution to create a number of question marks, but
this seems very dirty
or
finding someway to substitue tuples or lists into the statement - I'm
not sure if this should be done using Python or SQLite substitution
though.


I do this kind of thing sometimes:

<test.csv>
a,b,c
1,2,3
4,5,6
</test.csv>

<code>
import csv
import sqlite3

reader = csv.reader (open ("test.csv", "rb"))
csv_colnames = reader.next ()

db = sqlite3.connect (":memory:")
coldefs = ", ".join ("%s VARCHAR (200)" % c for c in csv_colnames)
db.execute ("CREATE TABLE test (%s)" % coldefs)

insert_cols = ", ".join (csv_colnames)
insert_qmarks = ", ".join ("?" for _ in csv_colnames)
insert_sql = "INSERT INTO test (%s) VALUES (%s)" % (insert_cols,
insert_qmarks)

db.executemany (insert_sql, list (reader))
for row in db.execute ("SELECT * FROM test"):
print row

</code>

Obviously, this is a proof-of-concept code. I'm (ab)using
the convenience functions at database level, I'm hardcoding
the column definitions, and I'm making a few other assumptions, but I
think it serves as an illustration. [..]
My code would probably look very similar. Btw you don't need to use
list() on an iterable to pass to executemany(). pysqlite's executemany()
accepts anything iterable (so generators work fine, too).

Also, with SQLite you can just skip data type definitions like
VARCHAR(200). They're ignored anyway.

-- Gerhard

Jul 30 '08 #3
Gerhard Häring wrote:
My code would probably look very similar. Btw you don't need to use
list() on an iterable to pass to executemany(). pysqlite's executemany()
accepts anything iterable (so generators work fine, too).
Thanks for that. My finger-memory told me to do that, possibly
because some *other* dbapi interface only accepts lists. Can't
quite remember. I'm usually all in favour of non-crystallised
iterators.
Also, with SQLite you can just skip data type definitions like
VARCHAR(200). They're ignored anyway.
Heh. Once again, finger memory forced me to put *something*
in there. I've been developing Enterprise databases for too
long :)

TJG
Jul 30 '08 #4
Hi and thanks,

I was hoping to avoid having to weld qmarks together but I guess
that's why people use things like SQL alchemy instead. It's a good
lesson anyway.

Thanks, Marc.
On Jul 30, 2:24*pm, Tim Golden <m...@timgolden.me.ukwrote:
Gerhard Häring wrote:
My code would probably look very similar. Btw you don't need to use
list() on an iterable to pass to executemany(). pysqlite's executemany()
accepts anything iterable (so generators work fine, too).

Thanks for that. My finger-memory told me to do that, possibly
because some *other* dbapi interface only accepts lists. Can't
quite remember. I'm usually all in favour of non-crystallised
iterators.
Also, with SQLite you can just skip data type definitions like
VARCHAR(200). They're ignored anyway.

Heh. Once again, finger memory forced me to put *something*
in there. I've been developing Enterprise databases for too
long :)

TJG
Jul 30 '08 #5

This discussion thread is closed

Replies have been disabled for this discussion.

Similar topics

19 posts views Thread by Skybuck Flying | last post: by
12 posts views Thread by John Salerno | last post: by
4 posts views Thread by Jim Carlock | last post: by
10 posts views Thread by Luigi | last post: by
9 posts views Thread by Gilles Ganault | last post: by
reply views Thread by Joe Goldthwaite | last post: by
20 posts views Thread by timotoole | last post: by
4 posts views Thread by guiromero | last post: by
By using this site, you agree to our Privacy Policy and Terms of Use.