python-sybase-misc

[python-sybase] problems with executing dynamically created queries.

[Erik T. <er...@th...>]

hi.

i am using 

- gentoo linux.
- sybase-0.36
- freetds-0.62.3
- python 2.3.3

from this linux machine i am accessing a mssql 7 server on an nt4.0 windows server.

i am doing

db=Sybase.connect('myserver','myuser','mypass','mydb')
c=db.cursor()
c.execute('select * from foo where something=@x', {'@x' : "blabla"})

it sais:

Sybase.DatabaseError: ------
Must declare the variable '@x'

??? i do not understand. why is this so?

i am trying to use python DB API 2.0. but i find the documentation for this
not very sufficient. there seems to be no good example at all on the net.

questions never covered and done wrong in almost all examples is the issue
with escaping the datatypes. some really do
c.execute("select from where x="%s"" % mystring)
this is of course fatal since mystring may contain closing " and then sql commands.

anyway... how do i get the module to work?


and... why is there several styles in DB API 2.0? i mean the Sybase.paramstyle
thing. i thought DB API was there for portability, and then i find out
that different modules have different ways to create sql queries. so there
is absolute zero portability. why?

thx & byebye
Erik


-- 
Erik Thiele

[python-sybase] Bug in param replacing in query (test case attached)

[Joshua Moore-O. <jo...@ch...>]

Basically, whenever I insert a text field into a table, it gets a whitespace added to the end..

I am using Sybase 0.36 with freetds

Here is a simple python script displaying this behaviour (server names and passwords removed for security)

import Sybase

sServer=''
sUser=''
sPassword=''
sDbName=''

conn = Sybase.connect( sServer, sUser, sPassword, sDbName )

cur = conn.cursor()

cur.execute( "CREATE TABLE tblTest( iId int, sIp text );" )

sSql = ( 'INSERT INTO tblTest ( iId, sIp )\n'
         '  VALUES( 1, @sIp ); \n' )

cur.execute( sSql
           ,  { '@sIp' : '12.34.56.87' } )

After this scripts runs, run this command

SELECT * FROM tblTest WHERE sIp LIKE '% ';

you will see the row come up..  and just as further proof that it's not a weird LIKE operator..

UPDATE tblTest SET sIp = rtrim(CAST( sIp AS varchar ));

will have show that it did indeed remove a space from the end as shown by a consecutive

SELECT * FROM tblTest WHERE sIp LIKE '% ';

Josh.

Re: [python-sybase] Pulling my hair out about a query... (solved?)

[Dave C. <dj...@ob...>]

Dave Cole wrote:
> Gregory Bond wrote:
> 
>> We still have 11.5 (youch!!), 12.0 and 12.5 servers on Solaris we can 
>> test
>> against. 
>> And if you ever get the bulkcopy implemented again we'd use it live!
> 
> 
> Ben keeps tellng me to reimplement the bulkcopy stuff.  If you are happy 
> to test it I should be able to do it over the weekend.

Have been sick for the last two days.  Getting bored at home so I had a 
crack at implementing bulkcopy again...

Install the attached Sybase.py and try something like the following:

import csv
import Sybase

db = Sybase.connect(server, user, passwd, database,
                     delay_connect=1, auto_commit=1)
db.set_property(Sybase.CS_BULK_LOGIN, Sybase.CS_TRUE)
db.connect()
blk = db.bulkcopy(table)
for row in csv.reader(file("some.csv")):
     blk.rowxfer(row)
blk.done()

- Dave

-- 
http://www.object-craft.com.au

Re: [python-sybase] Pulling my hair out about a query... (solved?)

[Dave C. <dj...@ob...>]

Gregory Bond wrote:
> We still have 11.5 (youch!!), 12.0 and 12.5 servers on Solaris we can test
> against.  
> 
> And if you ever get the bulkcopy implemented again we'd use it live!

Ben keeps tellng me to reimplement the bulkcopy stuff.  If you are happy 
to test it I should be able to do it over the weekend.

- Dave

-- 
http://www.object-craft.com.au

Re: [python-sybase] Pulling my hair out about a query... (solved?)

[Gregory B. <gn...@it...>]

We still have 11.5 (youch!!), 12.0 and 12.5 servers on Solaris we can test
against.  

And if you ever get the bulkcopy implemented again we'd use it live!

Re: [python-sybase] Pulling my hair out about a query... (solved?)

[Skip M. <sk...@po...>]

Dave & Stan,

I seem to have reached a useful conclusion about my query involving dates.
Our production Sybase server here is running 11.9.something while the Sybase
module is built using the 12.5 OCS libraries.  We have a test server running
12.5.  Grasping at straws this morning I tried running the script against
the test server.  Voila!  Valid results were returned!

That seems to solve my problem.  We will transition the test server to
production in a couple weeks and in the meantime I can just use the test
server directly.  That doesn't solve the problem for the Sybase module
though.  I suppose it's possible that some bit is set incompatibly at the
lower levels.  I'd be happy to try some more debugging if you want.  Just
let me know.

Thanks for the help,

-- 
Skip Montanaro
sk...@po...

Re: [python-sybase] Pulling my hair out about a query...

[Skip M. <sk...@po...>]

    Dave> Ooops.  I was confusing the Sybase module with another one of my 
    Dave> modules...  Should have checked the code.  Try this:

    Dave> import Sybase

    Dave> Sybase._ctx.debug = 1
    Dave> Sybase.set_debug(open('debug.log', 'w'))
    Dave>      :

Thanks, that worked much better.  Attached is the debug.log file.  I don't
see anything obvious, but I'm not really sure what I'm looking at either...

Skip

Re: [python-sybase] Pulling my hair out about a query...

[Dave C. <dj...@ob...>]

Skip Montanaro wrote:
>     Dave> Can you try doing something like the following:
> 
>     Dave> import Sybase
> 
>     Dave> Sybase.set_debug(open('debug.log', 'w'))
>     Dave>     :
>     Dave> conn = Sybase.Connection(..., debug=1)
>     Dave>     :
> 
> Dave,
> 
> Thanks for the response.  Alas, there is no debug arg to the Connection
> class constructor.  (I'm using 0.36.)  With just the set_debug() call
> nothing turns up in debug.log.  Is the debug argument something that's only
> available from CVS?

Ooops.  I was confusing the Sybase module with another one of my 
modules...  Should have checked the code.  Try this:

import Sybase

Sybase._ctx.debug = 1
Sybase.set_debug(open('debug.log', 'w'))
     :

- Dave

-- 
http://www.object-craft.com.au

Re: [python-sybase] Pulling my hair out about a query...

[Skip M. <sk...@po...>]

    Dave> Can you try doing something like the following:

    Dave> import Sybase

    Dave> Sybase.set_debug(open('debug.log', 'w'))
    Dave>     :
    Dave> conn = Sybase.Connection(..., debug=1)
    Dave>     :

Dave,

Thanks for the response.  Alas, there is no debug arg to the Connection
class constructor.  (I'm using 0.36.)  With just the set_debug() call
nothing turns up in debug.log.  Is the debug argument something that's only
available from CVS?

Skip

Re: [python-sybase] Pulling my hair out about a query...

[Dave C. <dj...@ob...>]

Skip Montanaro wrote:
>     >> I have a fairly straightforward query:
>     ...
>     >> that fails (returns no result rows) from Python. 
> 
>     stan> a.  Is it possible to get an error code from the
>     stan> cursor?  It should be an attribute you can print.
> 
> I don't see anything like an error attribute and cursor objects done have
> __getattr__ methods so it can't be a dynamically retrieved thing.  I also
> checked the cursor's _fetcher attribute and the connection for error-like
> attributes but didn't see anything.
> 
>     stan> b.  Can it be that the date routine is not being
>     stan> called until after the main query has run?  You could
>     stan> check this by declaring variables and setting them
>     stan> with the date call before you run the actual query. 
>     stan> If the dates are unassigned while the query runs, you
>     stan> would get no results.
> 
> I'm not sure what you mean here.  When run, this code
> 
>     import Sybase
> 
>     q = """\
>     SELECT trade_time, price
>       FROM underlying_daily_fp
>       WHERE symbol = @sym
>         AND instrument = @inst
>         AND trade_time BETWEEN @start AND @end
>         AND (price_type='H' or price_type='L')
>       ORDER BY trade_time DESC
>     """
> 
>     args = {'@sym': 'ES',
>             '@start': Sybase.Date(2004, 4, 19),
>             '@end': Sybase.Date(2004, 5, 17),
>             '@inst': 'F'
>             }
> 
>     print [(k,type(v),v) for (k,v) in args.items()]
> 
>     conn = Sybase.Connection(...)
>     c = conn.cursor()
>     c.execute(q, args)
>     print c.fetchall()
> 
> prints this output:
> 
>     [('@sym', <type 'str'>, 'ES'), ('@start', <type 'DateTimeType'>, 'Apr 19
>     2004 12:00AM'), ('@end', <type 'DateTimeType'>, 'May 17 2004 12:00AM'),
>     ('@inst', <type 'str'>, 'F')] 
>     []
> 
> The DateTimeType objects are clearly created before the query executes.  In
> any case, I'm not sure what you mean by "declaring variables and setting
> them with the date call".  Is this something I embed in my query string?
> 
> If I twiddle the trade_time clause to be any of
> 
>         AND trade_time >= @start
>         AND trade_time <= @end
> 
> or
> 
>         AND trade_time >= @start
> 
> or
> 
>         AND trade_time <= @end
> 
> the query always returns empty results.  If I remove the trade_time clause
> altogether the query succeeds, but of course the date range of the returned
> results is incorrect.

Can you try doing something like the following:

import Sybase

Sybase.set_debug(open('debug.log', 'w'))
    :
conn = Sybase.Connection(..., debug=1)
    :

Then the module will produce quite a bit of debugging output in the log 
file.  I hope that some of the output in the log file will shed some 
light on what is going on.

- Dave

-- 
http://www.object-craft.com.au

Re: [python-sybase] Pulling my hair out about a query...

[Skip M. <sk...@po...>]

    >> I have a fairly straightforward query:
    ...
    >> that fails (returns no result rows) from Python. 

    stan> a.  Is it possible to get an error code from the
    stan> cursor?  It should be an attribute you can print.

I don't see anything like an error attribute and cursor objects done have
__getattr__ methods so it can't be a dynamically retrieved thing.  I also
checked the cursor's _fetcher attribute and the connection for error-like
attributes but didn't see anything.

    stan> b.  Can it be that the date routine is not being
    stan> called until after the main query has run?  You could
    stan> check this by declaring variables and setting them
    stan> with the date call before you run the actual query. 
    stan> If the dates are unassigned while the query runs, you
    stan> would get no results.

I'm not sure what you mean here.  When run, this code

    import Sybase

    q = """\
    SELECT trade_time, price
      FROM underlying_daily_fp
      WHERE symbol = @sym
        AND instrument = @inst
        AND trade_time BETWEEN @start AND @end
        AND (price_type='H' or price_type='L')
      ORDER BY trade_time DESC
    """

    args = {'@sym': 'ES',
            '@start': Sybase.Date(2004, 4, 19),
            '@end': Sybase.Date(2004, 5, 17),
            '@inst': 'F'
            }

    print [(k,type(v),v) for (k,v) in args.items()]

    conn = Sybase.Connection(...)
    c = conn.cursor()
    c.execute(q, args)
    print c.fetchall()

prints this output:

    [('@sym', <type 'str'>, 'ES'), ('@start', <type 'DateTimeType'>, 'Apr 19
    2004 12:00AM'), ('@end', <type 'DateTimeType'>, 'May 17 2004 12:00AM'),
    ('@inst', <type 'str'>, 'F')] 
    []

The DateTimeType objects are clearly created before the query executes.  In
any case, I'm not sure what you mean by "declaring variables and setting
them with the date call".  Is this something I embed in my query string?

If I twiddle the trade_time clause to be any of

        AND trade_time >= @start
        AND trade_time <= @end

or

        AND trade_time >= @start

or

        AND trade_time <= @end

the query always returns empty results.  If I remove the trade_time clause
altogether the query succeeds, but of course the date range of the returned
results is incorrect.

Thx,

-- 
Skip Montanaro
sk...@po...

Re: [python-sybase] Pulling my hair out about a query...

[stan l. <st...@ya...>]

--- Skip Montanaro <sk...@po...> wrote:
> I have a fairly straightforward query:
> 
>     SELECT trade_time, price
>       FROM underlying_daily_fp d, underlying_months
> m
>       WHERE d.instrument = @inst
>         AND d.symbol = @sym
>         AND m.instrument=d.instrument
>         AND m.symbol=d.symbol
>         AND (price_type='H' or price_type='L')
>         AND m.exp_year=d.exp_year
>         AND m.exp_month=d.exp_month
>         AND trade_time BETWEEN @start AND @end
>         AND d.trade_time=m.date
>         AND m.exp_num=1
>       ORDER BY trade_time  DESC
> 
> with args:
> 
>     {'@sym': 'ES',
>      '@start': 'Apr 19 2004 12:00AM',
>      '@end': 'May 17 2004 12:00AM',
>      '@inst': 'F'}
> 
> that fails (returns no result rows) from Python. 
> (The @start and @end
> params are actually Sybase.Date() objects.) 
> However, the equivalent code
> pasted into isql (using strings for dates) works as
> expected:
> 
>     1>  select trade_time, price from
> underlying_daily_fp d, underlying_months m
>        where d.instrument='F'
>        and d.symbol='ES'
>        and m.instrument=d.instrument
>      and m.symbol=d.symbol
>      and (price_type='H' or price_type='L')
>      AND m.exp_year=d.exp_year
>      AND m.exp_month=d.exp_month
>      AND trade_time BETWEEN 'Apr 19 2004 12:00AM'
> and 'May 17 2004 12:00AM'
>      AND d.trade_time=m.date
>      AND m.exp_num=1
>      ORDER BY trade_time  DESC
>     2> 3> 4> 5> 6> 7> 8> 9> 10> 11> 12> 13> 
>     14> go
>      trade_time                 price               
> 
>      -------------------------- --------------------
> 
>             May 17 2004 12:00AM          1078.250000
> 
>             May 17 2004 12:00AM          1089.250000
> 
>             ...
>             Apr 19 2004 12:00AM          1135.250000
> 
> 
>     (42 rows affected)
>     1> 
> 
> Furthermore, if I manually expand the arguments and
> call the connection's or
> cursor's execute() methods:
> 
>     >>> q
>     "SELECT trade_time, price\n  FROM
> underlying_daily_fp d, underlying_months m\n  WHERE
> d.instrument = 'F'\n    AND d.symbol = 'ES'\n    AND
> m.instrument=d.instrument\n    AND
> m.symbol=d.symbol\n    AND (price_type='H' or
> price_type='L')\n    AND m.exp_year=d.exp_year\n   
> AND m.exp_month=d.exp_month\n    AND trade_time
> BETWEEN 'Apr 19 2004 12:00AM' AND 'May 17 2004
> 12:00AM'\n    AND d.trade_time=m.date\n    AND
> m.exp_num=1\n  ORDER BY trade_time  DESC"
>     >>> len(db.execute(q)[0])
>     42
>     >>> c = db.cursor()
>     >>> c.execute(q)
>     >>> len(c.fetchall())
>     42
> 
> it also works.
> 
> Is it date handling?  Is it possible to get the
> actual SQL statement
> executed by the server so I can see if and where
> I've gone wrong?  Do other
> people experience this problem (works from isql but
> not Python)?  The
> platform is OCS 12.5 and the latest version of the
> Sybase module running on
> Mandrake 10.
> 
> Thanks,
> 
> Skip Montanaro
> sk...@po...
> _______________________________________________
> Python-sybase mailing list
> Pyt...@ww...
>
https://www.object-craft.com.au/cgi-bin/mailman/listinfo/python-sybase
> 


Hi Skip,

I don't have a solution to your problem, just a couple
of suggestions.  

a.  Is it possible to get an error code from the
cursor?  It should be an attribute you can print.
b.  Can it be that the date routine is not being
called until after the main query has run?  You could
check this by declaring variables and setting them
with the date call before you run the actual query. 
If the dates are unassigned while the query runs, you
would get no results.

Just some thoughts.

Stan



		
__________________________________
Do you Yahoo!?
New and Improved Yahoo! Mail - Send 10MB messages!
http://promotions.yahoo.com/new_mail 

[python-sybase] Pulling my hair out about a query...

[Skip M. <sk...@po...>]

I have a fairly straightforward query:

    SELECT trade_time, price
      FROM underlying_daily_fp d, underlying_months m
      WHERE d.instrument = @inst
        AND d.symbol = @sym
        AND m.instrument=d.instrument
        AND m.symbol=d.symbol
        AND (price_type='H' or price_type='L')
        AND m.exp_year=d.exp_year
        AND m.exp_month=d.exp_month
        AND trade_time BETWEEN @start AND @end
        AND d.trade_time=m.date
        AND m.exp_num=1
      ORDER BY trade_time  DESC

with args:

    {'@sym': 'ES',
     '@start': 'Apr 19 2004 12:00AM',
     '@end': 'May 17 2004 12:00AM',
     '@inst': 'F'}

that fails (returns no result rows) from Python.  (The @start and @end
params are actually Sybase.Date() objects.)  However, the equivalent code
pasted into isql (using strings for dates) works as expected:

    1>  select trade_time, price from underlying_daily_fp d, underlying_months m
       where d.instrument='F'
       and d.symbol='ES'
       and m.instrument=d.instrument
     and m.symbol=d.symbol
     and (price_type='H' or price_type='L')
     AND m.exp_year=d.exp_year
     AND m.exp_month=d.exp_month
     AND trade_time BETWEEN 'Apr 19 2004 12:00AM' and 'May 17 2004 12:00AM'
     AND d.trade_time=m.date
     AND m.exp_num=1
     ORDER BY trade_time  DESC
    2> 3> 4> 5> 6> 7> 8> 9> 10> 11> 12> 13> 
    14> go
     trade_time                 price                
     -------------------------- -------------------- 
            May 17 2004 12:00AM          1078.250000 
            May 17 2004 12:00AM          1089.250000 
            ...
            Apr 19 2004 12:00AM          1135.250000 

    (42 rows affected)
    1> 

Furthermore, if I manually expand the arguments and call the connection's or
cursor's execute() methods:

    >>> q
    "SELECT trade_time, price\n  FROM underlying_daily_fp d, underlying_months m\n  WHERE d.instrument = 'F'\n    AND d.symbol = 'ES'\n    AND m.instrument=d.instrument\n    AND m.symbol=d.symbol\n    AND (price_type='H' or price_type='L')\n    AND m.exp_year=d.exp_year\n    AND m.exp_month=d.exp_month\n    AND trade_time BETWEEN 'Apr 19 2004 12:00AM' AND 'May 17 2004 12:00AM'\n    AND d.trade_time=m.date\n    AND m.exp_num=1\n  ORDER BY trade_time  DESC"
    >>> len(db.execute(q)[0])
    42
    >>> c = db.cursor()
    >>> c.execute(q)
    >>> len(c.fetchall())
    42

it also works.

Is it date handling?  Is it possible to get the actual SQL statement
executed by the server so I can see if and where I've gone wrong?  Do other
people experience this problem (works from isql but not Python)?  The
platform is OCS 12.5 and the latest version of the Sybase module running on
Mandrake 10.

Thanks,

Skip Montanaro
sk...@po...

Re: [python-sybase] Can't put LIKE parameter in parameter dict?

[Marcos P. <ma...@bu...>]

It might be that the db server/api takes incorrect decisions about the
type of the parameter...

You may try casting the parameter:

 and symbol like cast(@s as varchar(100))

By the way, you might use sigle quotes around string constants, it is
more portable (triple quotes around sql looks better too ;-)

El jue, 10-06-2004 a las 15:18, Skip Montanaro escribi=F3:
> Can someone explain why these two seemingly identical queries return
> different results?  Can't LIKE wildcards be given in parameter dictionari=
es?
>=20
>     >>> c.execute('select distinct symbol, open_time, close_time,'
>     ...           '                session_start, session_end, instrument=
'
>     ...           '  from underlying_sessions'
>     ...           '  where instrument =3D @i'
>     ...           '    and symbol like "M%"',
>     ...           {"@i": "E"})
>     >>> len(c.fetchall())
>     402
>     >>> c.execute('select distinct symbol, open_time, close_time,'
>     ...           '                session_start, session_end, instrument=
'
>     ...           '  from underlying_sessions'
>     ...           '  where instrument =3D @i'
>     ...           '    and symbol like @s',
>     ...           {"@i": "E",
>     ...            "@s": "M%"})
>     >>> len(c.fetchall())
>     0
>=20
> Thanks,

[python-sybase] Can't put LIKE parameter in parameter dict?

[Skip M. <sk...@po...>]

Can someone explain why these two seemingly identical queries return
different results?  Can't LIKE wildcards be given in parameter dictionaries?

    >>> c.execute('select distinct symbol, open_time, close_time,'
    ...           '                session_start, session_end, instrument'
    ...           '  from underlying_sessions'
    ...           '  where instrument = @i'
    ...           '    and symbol like "M%"',
    ...           {"@i": "E"})
    >>> len(c.fetchall())
    402
    >>> c.execute('select distinct symbol, open_time, close_time,'
    ...           '                session_start, session_end, instrument'
    ...           '  from underlying_sessions'
    ...           '  where instrument = @i'
    ...           '    and symbol like @s',
    ...           {"@i": "E",
    ...            "@s": "M%"})
    >>> len(c.fetchall())
    0

Thanks,

-- 
Skip Montanaro
Got gigs? http://www.musi-cal.com/submit.html
Got spam? http://www.spambayes.org/
sk...@po...

RE: [python-sybase] Python Sybase-0.36 module with SUSE Linux 9.0

[Skip M. <sk...@po...>]

    Markus> I am still stuck with what seems to be an Sybase error though:

    >>>> db = Sybase.connect('LINUX', 'sa', 'XXX')
    >>>> c=db.cursor()
    >>>> c.callproc('sp_help')
    Markus> Traceback (most recent call last):
    Markus>   File "<stdin>", line 1, in ?
    Markus>   File "/usr/lib/python2.3/site-packages/Sybase.py", line 653, in callproc
    Markus>     self.description = fetcher.start(self.arraysize, out_params)
    Markus>   File "/usr/lib/python2.3/site-packages/Sybase.py", line 338, in start
    Markus>     return _FetchNow.start(self, arraysize)
    Markus>   File "/usr/lib/python2.3/site-packages/Sybase.py", line 262, in start
    Markus>     self._raise_error(Error, 'ct_send')
    Markus>   File "/usr/lib/python2.3/site-packages/Sybase.py", line 287, in _raise_error
    Markus>     raise exc(text)
    Markus> Sybase.Error: ct_send

I don't recognize that.  Do you also get it with something simple like

    c.execute("select 1")
    print c.fetchall()

?

Skip

RE: [python-sybase] Python Sybase-0.36 module with SUSE Linux 9.0

[<m.d...@BG...>]

Hi Skip,

thanks a lot!
Works immediately and a lot easier than using ldconfig.

I am still stuck with what seems to be an Sybase error though:

>>> db =3D Sybase.connect('LINUX', 'sa', 'XXX')
>>> c=3Ddb.cursor()
>>> c.callproc('sp_help')
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "/usr/lib/python2.3/site-packages/Sybase.py", line 653, in =
callproc
    self.description =3D fetcher.start(self.arraysize, out_params)
  File "/usr/lib/python2.3/site-packages/Sybase.py", line 338, in start
    return _FetchNow.start(self, arraysize)
  File "/usr/lib/python2.3/site-packages/Sybase.py", line 262, in start
    self._raise_error(Error, 'ct_send')
  File "/usr/lib/python2.3/site-packages/Sybase.py", line 287, in =
_raise_error
    raise exc(text)
Sybase.Error: ct_send



thanks,
Markus


-----Original Message-----
From: Skip Montanaro [mailto:sk...@po...]
Sent: Mittwoch, 9. Juni 2004 15:00
To: D=F6ring, Markus
Cc: pyt...@ww...
Subject: Re: [python-sybase] Python Sybase-0.36 module with SUSE Linux
9.0


Try running=20

    python setup.py build_ext --help

There is a --rpath flag you can give when building extensions.  I think
something like

    python setup.py clean -a
    python setup.py build_ext --rpath=3D$SYBASE_OCS/lib
    python setup.py install

would probably do the trick.

--=20
Skip Montanaro
Got gigs? http://www.musi-cal.com/submit.html
Got spam? http://www.spambayes.org/
sk...@po...

[python-sybase] RE: Python Sybase-0.36 module with SUSE Linux 9.0

[<m.d...@BG...>]

Hello,

I have found that when I manually force the dynamic linker to include =
the sybase OCS library directory with ldconfig it works!

It worked like this:

linux:~/sybase-0.36 # ldconfig -p | grep sybase
linux:~/sybase-0.36 # python setup.py build_ext
linux:~/sybase-0.36 # ldd build/lib.linux-i686-2.3/sybasect.so
        libct.so =3D> /usr/local/sybase/OCS-12_5/lib/libct.so =
(0x4003d000)
        libcs.so =3D> /usr/local/sybase/OCS-12_5/lib/libcs.so =
(0x400ab000)
        libsybtcl.so =3D> /usr/local/sybase/OCS-12_5/lib/libsybtcl.so =
(0x400b8000)
        libcomn.so =3D> /usr/local/sybase/OCS-12_5/lib/libcomn.so =
(0x400f0000)
        libintl.so =3D> /usr/local/sybase/OCS-12_5/lib/libintl.so =
(0x40154000)
        libpthread.so.0 =3D> /lib/i686/libpthread.so.0 (0x4015e000)
        libc.so.6 =3D> /lib/i686/libc.so.6 (0x401af000)
        /lib/ld-linux.so.2 =3D> /lib/ld-linux.so.2 (0x80000000)


et voila!
Hopefully this build works now...

But is this the way to do it right?


Thanks,
Markus



-----Original Message-----
From: D=F6ring, Markus=20
Sent: Mittwoch, 9. Juni 2004 13:16
To: 'pyt...@ww...'
Subject: Python Sybase-0.36 module with SUSE Linux 9.0


Hi,

I am trying to get the sybase 0.36 module running on SUSE Linux 9.0 with =
ASE 12.5 developer version installed in /usr/local/sybase

I managed to compile the shared libraries and run python setup.py =
install by setting the proper environment variables for SYBASE, =
SYBASE_OCS and SYBASE_ASE.

But it seems the sybasect.so file is not linked correctly.
When I test the build file with ldd, I get this:

----------------------------------------
linux:~/sybase-0.36 # ldd build/lib.linux-i686-2.3/sybasect.so
        libct.so =3D> not found
        libcs.so =3D> not found
        libsybtcl.so =3D> not found
        libcomn.so =3D> not found
        libintl.so =3D> not found
        libpthread.so.0 =3D> /lib/i686/libpthread.so.0 (0x4003e000)
        libc.so.6 =3D> /lib/i686/libc.so.6 (0x4008f000)
        /lib/ld-linux.so.2 =3D> /lib/ld-linux.so.2 (0x80000000)


----------------------------------------

the gcc build gives me this:
----------------------------------------
gcc -pthread -fno-strict-aliasing -DNDEBUG -D_FILE_OFFSET_BITS=3D64 =
-DHAVE_LARGEFILE_SUPPORT -O2 -march=3Di586 -mcpu=3Di686 =
-fmessage-length=3D0 -fPIC -Wall -fPIC -DWANT_BULKCOPY -DHAVE_BLK_ALLOC =
-DHAVE_BLK_DESCRIBE -DHAVE_BLK_DROP -DHAVE_BLK_ROWXFER_MULT =
-DHAVE_BLK_TEXTXFER -DHAVE_CT_CURSOR -DHAVE_CT_DATA_INFO =
-DHAVE_CT_DYNAMIC -DHAVE_CT_SEND_DATA -DHAVE_CT_SETPARAM -DHAVE_CS_CALC =
-DHAVE_CS_CMP -I/usr/local/sybase/OCS-12_5/include =
-I/usr/include/python2.3 -c sybasect.c -o =
build/temp.linux-i686-2.3/sybasect.o
...
gcc -pthread -shared build/temp.linux-i686-2.3/blk.o =
build/temp.linux-i686-2.3/databuf.o build/temp.linux-i686-2.3/cmd.o =
build/temp.linux-i686-2.3/conn.o build/temp.linux-i686-2.3/ctx.o =
build/temp.linux-i686-2.3/datafmt.o build/temp.linux-i686-2.3/iodesc.o =
build/temp.linux-i686-2.3/locale.o build/temp.linux-i686-2.3/msgs.o =
build/temp.linux-i686-2.3/numeric.o build/temp.linux-i686-2.3/money.o =
build/temp.linux-i686-2.3/datetime.o =
build/temp.linux-i686-2.3/sybasect.o -L/usr/local/sybase/OCS-12_5/lib =
-lblk -lct -lcs -lsybtcl -lcomn -lintl -o =
build/lib.linux-i686-2.3/sybasect.so
running install_lib
copying build/lib.linux-i686-2.3/sybasect.so -> =
/usr/lib/python2.3/site-packages



Does anyone have an idea how to solve this?
Any help greatly appreciated.

Thanks,
Markus
--
 Markus D=F6ring=20
 Botanic Garden and Botanical Museum Berlin Dahlem,=20
 Dept. of Biodiversity Informatics=20
 K=F6nigin-Luise-Str. 6-8, D-14191 Berlin=20
 Phone: +49 30 83850-284
 Email: m.d...@bg...=20

Re: [python-sybase] Python Sybase-0.36 module with SUSE Linux 9.0

[Skip M. <sk...@po...>]

    Markus> When I test the build file with ldd, I get this:

    Markus> ----------------------------------------
    Markus> linux:~/sybase-0.36 # ldd build/lib.linux-i686-2.3/sybasect.so
    Markus>         libct.so => not found
    Markus>         libcs.so => not found
    Markus>         libsybtcl.so => not found
    Markus>         libcomn.so => not found
    Markus>         libintl.so => not found
    Markus>         libpthread.so.0 => /lib/i686/libpthread.so.0 (0x4003e000)
    Markus>         libc.so.6 => /lib/i686/libc.so.6 (0x4008f000)
    Markus>         /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x80000000)

    ...

    Markus> Does anyone have an idea how to solve this?
    Markus> Any help greatly appreciated.

Try running 

    python setup.py build_ext --help

There is a --rpath flag you can give when building extensions.  I think
something like

    python setup.py clean -a
    python setup.py build_ext --rpath=$SYBASE_OCS/lib
    python setup.py install

would probably do the trick.

-- 
Skip Montanaro
Got gigs? http://www.musi-cal.com/submit.html
Got spam? http://www.spambayes.org/
sk...@po...

[python-sybase] Python Sybase-0.36 module with SUSE Linux 9.0

[<m.d...@BG...>]

Hi,

I am trying to get the sybase 0.36 module running on SUSE Linux 9.0 with =
ASE 12.5 developer version installed in /usr/local/sybase

I managed to compile the shared libraries and run python setup.py =
install by setting the proper environment variables for SYBASE, =
SYBASE_OCS and SYBASE_ASE.

But it seems the sybasect.so file is not linked correctly.
When I test the build file with ldd, I get this:

----------------------------------------
linux:~/sybase-0.36 # ldd build/lib.linux-i686-2.3/sybasect.so
        libct.so =3D> not found
        libcs.so =3D> not found
        libsybtcl.so =3D> not found
        libcomn.so =3D> not found
        libintl.so =3D> not found
        libpthread.so.0 =3D> /lib/i686/libpthread.so.0 (0x4003e000)
        libc.so.6 =3D> /lib/i686/libc.so.6 (0x4008f000)
        /lib/ld-linux.so.2 =3D> /lib/ld-linux.so.2 (0x80000000)


----------------------------------------

the gcc build gives me this:
----------------------------------------
gcc -pthread -fno-strict-aliasing -DNDEBUG -D_FILE_OFFSET_BITS=3D64 =
-DHAVE_LARGEFILE_SUPPORT -O2 -march=3Di586 -mcpu=3Di686 =
-fmessage-length=3D0 -fPIC -Wall -fPIC -DWANT_BULKCOPY -DHAVE_BLK_ALLOC =
-DHAVE_BLK_DESCRIBE -DHAVE_BLK_DROP -DHAVE_BLK_ROWXFER_MULT =
-DHAVE_BLK_TEXTXFER -DHAVE_CT_CURSOR -DHAVE_CT_DATA_INFO =
-DHAVE_CT_DYNAMIC -DHAVE_CT_SEND_DATA -DHAVE_CT_SETPARAM -DHAVE_CS_CALC =
-DHAVE_CS_CMP -I/usr/local/sybase/OCS-12_5/include =
-I/usr/include/python2.3 -c sybasect.c -o =
build/temp.linux-i686-2.3/sybasect.o
...
gcc -pthread -shared build/temp.linux-i686-2.3/blk.o =
build/temp.linux-i686-2.3/databuf.o build/temp.linux-i686-2.3/cmd.o =
build/temp.linux-i686-2.3/conn.o build/temp.linux-i686-2.3/ctx.o =
build/temp.linux-i686-2.3/datafmt.o build/temp.linux-i686-2.3/iodesc.o =
build/temp.linux-i686-2.3/locale.o build/temp.linux-i686-2.3/msgs.o =
build/temp.linux-i686-2.3/numeric.o build/temp.linux-i686-2.3/money.o =
build/temp.linux-i686-2.3/datetime.o =
build/temp.linux-i686-2.3/sybasect.o -L/usr/local/sybase/OCS-12_5/lib =
-lblk -lct -lcs -lsybtcl -lcomn -lintl -o =
build/lib.linux-i686-2.3/sybasect.so
running install_lib
copying build/lib.linux-i686-2.3/sybasect.so -> =
/usr/lib/python2.3/site-packages



Does anyone have an idea how to solve this?
Any help greatly appreciated.

Thanks,
Markus
--
 Markus D=F6ring=20
 Botanic Garden and Botanical Museum Berlin Dahlem,=20
 Dept. of Biodiversity Informatics=20
 K=F6nigin-Luise-Str. 6-8, D-14191 Berlin=20
 Phone: +49 30 83850-284
 Email: m.d...@bg...=20

[python-sybase] More Pythonic default args

[Skip M. <sk...@po...>]

It seems more Pythonic to use None as a default parameter value instead of
mutable objects like dictionaries and lists, even when the parameters aren't
modified by the function.  A simple patch for execute() and executemany() is
attached.

-- 
Skip Montanaro
Got gigs? http://www.musi-cal.com/submit.html
Got spam? http://www.spambayes.org/
sk...@po...

[python-sybase] Limitation in parameter key length?

[Skip M. <sk...@po...>]

Can someone explain why two select statements should yield different
results?

    >>> c.execute('select instrument,point_value,currency,tick_size,min_price_change'
    ...                   '  from history..underlying_contract_data_fp'
    ...                   '  where symbol = @symbol' 
    ...                   '    and instrument = @instrument',
    ...            {"@symbol": "MSFT", "@instrument": "E"})
    >>> c.fetchall()
    []
    >>> c.execute('select instrument,point_value,currency,tick_size,min_price_change'
    ...                   '  from history..underlying_contract_data_fp'
    ...                   '  where symbol = @sym'
    ...                   '    and instrument = @inst',
    ...            {"@sym": "MSFT", "@inst": "E"})
    >>> c.fetchall()
    [('E', 1.0, 'US    ', 0.01, 0.01)]

It seems to be a bug in the Sybase module's parameter expansion.  Is there
some limitation on the length of a parameter key?  This certainly suggests
there is:

    >>> c.execute('select instrument,point_value,currency,tick_size,min_price_change'
    ...                   '  from history..underlying_contract_data_fp'
    ...                   '  where symbol = @sym'
    ...                   '    and instrument = @instrume',
    ...            {"@sym": "MSFT", "@instrume": "E"})
    >>> c.fetchall()
    []
    >>> c.execute('select instrument,point_value,currency,tick_size,min_price_change'
    ...                   '  from history..underlying_contract_data_fp'
    ...                   '  where symbol = @sym'
    ...                   '    and instrument = @instrum',
    ...            {"@sym": "MSFT", "@instrum": "E"})
    >>> c.fetchall()
    [('E', 1.0, 'US    ', 0.01, 0.01)]

If there is, it should probably be documented.

-- 
Skip Montanaro
Got gigs? http://www.musi-cal.com/submit.html
Got spam? http://www.spambayes.org/
sk...@po...

[python-sybase] Re: Thank you!

[Harri.pasanen <har...@tr...>]

<html><body>
  

<br>
</body></html>

[python-sybase] Incoming message

[Harri.pasanen <har...@tr...>]

<html><body>
 

<br>
</body></html>

[python-sybase] Re: Thanks :)

[Harri.pasanen <har...@tr...>]

<html><body>
 

<br>
</body></html>