For the same task I used a custom factor returning the sid and created a filter with it.
Actually I hoped something like that would be builtin but I couldn't find it in the single-mega-ultra-combined-help-tutorial-reference page.

Here's what I mean, the only comment is the bit you asked for, the trick is that the asset parameter to compute() holds all the sids for the pipeline securities:

from quantopian.algorithm import attach_pipeline, pipeline_output  
from quantopian.pipeline import Pipeline, CustomFactor

class SidFactor(CustomFactor):  
    inputs = []  
    window_length = 1

    def compute(self, today, assets, out):  
        out[:] = assets

def initialize(context):  
    sid = SidFactor()

    # This is the filter you want  
    my_sid_filter = sid.eq(24)

    pipe = Pipeline()  
    pipe.add(sid, 'sid')  
    pipe.set_screen(my_sid_filter)  
    attach_pipeline(pipe, 'just_my_sid')     

def handle_data(context, data):  
    print(pipeline_output('just_my_sid'))

Thank you so much Andrea, it worked like a charm!
Is it possible to reference it to a list of sids instead of just one?
I'm trying to filter it to show today's open positions.

Try this (I am not sure how fast it can be though):

class SidFactor(CustomFactor):  
    inputs = []  
    window_length = 1

    def compute(self, today, assets, out):  
        out[:] = assets

def create_sid_screen(sid_list):  
    sid_factor = SidFactor()  
    sid_screen = None  
    for s in sid_list:  
        sid_screen = sid_factor.eq(s) if (sid_screen is None) else (sid_screen | sid_factor.eq(s))  
    return sid_screen

sid_list = [1,2,3,4,,5,6,7,8,9....,666]

sid_screen = create_sid_screen(sid_list)

pipe.set_screen(sid_screen)  

Months ago I asked to introduce a feature complementary to this one in Pipeline but no answer yet.

A hack because Quantopian doesn't allow passing arguments to compute(), but it works. There's a note on the numpy in1d documentation that says something along the lines of 'in1d is slow. Use Pandas instead.' But that's left as an exercise for the reader.

from quantopian.algorithm import attach_pipeline, pipeline_output  
from quantopian.pipeline import Pipeline, CustomFactor  
import numpy as np

class SidFactor(CustomFactor):  
    inputs = []  
    window_length = 1  
    sids = []

    def compute(self, today, assets, out):  
        out[:] = np.in1d(assets, self.sids)

def initialize(context):  
    sid = SidFactor()  
    sid.sids = (19662, 40107)

    pipe = Pipeline(  
        columns={ 'sid': sid },  
        screen=sid.eq(1))

    attach_pipeline(pipe, 'pipe')     

def handle_data(context, data):  
    print(pipeline_output('pipe'))  

Hey guys, these are great workarounds! I've submitted an internal feature request for a Pipeline filter for a list of SIDs.

Hi, Zipline has a Pipeline filter for a list of static assets called: StaticAssets in: zipline.pipeline.filters.filter.
I am writing this over half a year after the last post, so I am not sure if StaticAssets existed before this thread or was a product of this thread.

Anyways you can go from a list of sids or tickers to assets then use this filter, it looks like it is available on Quantopian under quantopian.pipeline.filters.

Indeed. It appears that StaticAssets was added to zipline on Nov 1. Much cleaner than the hacks that luca and I did.

btw, syntax is this:

from quantopian.pipeline.filters import  StaticAssets

aapl = StaticAssets(symbols(['AAPL']))  
aapl_ibm = StaticAssets(symbols(['AAPL', 'IBM']))  

Thanks to let everybody knows, that is a good addition to pipeline indeed.