Using parsestream blocks until fully read

[resamsel]

I was trying to combine sqlparse.parsestream with reading from stdin, but could not get it to work properly. Here's my approach:

stream.py:

import sys
import os
import sqlparse

#newin = os.fdopen(sys.stdin.fileno(), 'r', 1)

s = sqlparse.parsestream(sys.stdin)

for i in s:
    print i

Calling it with:

for i in `seq 1 5`; do echo select $i\;; sleep 1; done | python stream.py

The output becomes available after 5 seconds. I was expecting parsestream to output each statement as soon as it is read from stdin. Also, I tried to create a separate newin that uses a buffersize of 1.

Is there anything I'm doing wrong here?

I'm in OS X Yosemite with Python 2.7.9.

[andialbrecht]

That's an interesting use case! ATM parsestream relies on EOF and is intended to work with patterns like "echo 'select * from foo' | stream.py". Are you trying to format log files in a tail-like manner?

For now parsestream doesn't try to deteremine the end of an statement at all (albeit it might be possible). This would be required for that use case then.

[resamsel]

Thank's for you reply! I'm having a large file with lots of SQL statements (70K+ lines) that I want to start as soon as the first statement has been read successfully. At the same time I want to be able to create SQL statements with a script and pipe it to my stream.py.

I already use sqlparse to read SQL statements from a file/stdin and execute them (see https://github.com/resamsel/dbnavigator/blob/master/src/dbnav/executer/__init__.py if you're interested in what I'm currently working on). I'm trying to replace the read_statements function with a stream, to work around the long time it takes to read and parse all of those 70K+ lines...

[mehaase]

+1 I'm interested in reading a database dump (.sql) and being able to extract information from Python without needing to actually stage it in a real database. I'm working with files > 100k lines. The performance is okay so far, but as I scale up, it would be nice if the generator could start returning results without needing to read the entire file into memory.

[c3c]

+1 i would like to use this library for parsing large sql dump files

[vmuriart]

will anyone be able to provide an example and test for this? I think I see what's going on, but would like to validate

[resamsel]

@vmuriart, what's wrong with the example and test in the original post? I'd be happy to provide more information, if you need any.

[vmuriart]

I'm on a Windows computer :sob:. Specifically though, I was looking for something that could be added to the libraries tests

[mehaase]

Here are some free SQL dumps found via Google. Some are quite large.

http://dev.mysql.com/doc/index-other.html http://sportsdb.org/sd/samples More: https://www.google.com/search?safe=off&q=filetype%3Asql

[vmuriart]

Thanks @mehaase. I meant an example that can be included on the tests module. I am having a hard time coming up with a way to automate this test.

[abudis]

Sorry for reviving an old issue, but it would be great if something similar (exactly the same?) as what @Toilal has added in the fork was added upstream!