On Wednesday, March 9, 2016 at 3:17:26 PM UTC-5,
jbrow...@clearcosthealth.com wrote:
>
> Like I said I am not terribly network savvy but what I meant was that the
> particular vpn client that I am using says that my "connection speed" is
> ~540Mpbs.
>
That's most-likely a speed to your router
Like I said I am not terribly network savvy but what I meant was that the
particular vpn client that I am using says that my "connection speed" is
~540Mpbs.
I know comparing sql alchemy to raw sql is kind of a no-no but I was just
illustrating that the there isn't obvious evidence of a slow net
Those aren't the same inserts as SqlAlchemy
The SQL you noted above (DECLARE / BEGIN/ END) is one packet of text that
is sent over the wire to the database server and executed. (a few hundred
bytes)
A more appropriate comparison would be sending 1000 separate "INSERT INTO
" statements ove
On 03/09/2016 11:56 AM, jbrownst...@clearcosthealth.com wrote:
This is all extremely helpful.
After some hair pulling I was able to get the local sql server express
running and I ran the same test on a local instance.
10100.1260.0000.1260.000 {method 'execute' of
'pyodbc
This is all extremely helpful.
After some hair pulling I was able to get the local sql server express
running and I ran the same test on a local instance.
10100.1260.0000.1260.000 {method 'execute' of
'pyodbc.Cursor' objects}
10100.0600.0000.0600.000 {m
so the profile shows pretty clearly that the time is spent on the DB
side. 1006 SQL statements are clocking at .09 seconds each (!) to take
a total of 93 seconds:
10060.0120.000 93.8730.093
C:\Users\jbrownstein\Documents\GitHub\ETL\ETL\env\lib\site-packages\sqlalchemy\engine\b
I'm on sqla 1.0.12 / MS SQL Server 2014 / python 64bit 2.7 (sorry should
have included that in the first post)
Here is an extract from the profile using the code
from http://docs.sqlalchemy.org/en/rel_0_8/faq.html#code-profiling
I am reading explanations in the FAQ now but any guidance would be
Thanks Mike - will the full profile ASAP.
On Monday, March 7, 2016, Mike Bayer wrote:
> What will make inserts slow is network as well as indexes. If you aren't
> comparing your script to some other backend then there's probably a whole
> lot of other things that can get in the way like unicod
What will make inserts slow is network as well as indexes. If you aren't
comparing your script to some other backend then there's probably a whole lot
of other things that can get in the way like unicode encoding, large columns or
other surprises in the model. If you run the full profile ver
I've run sqlalchemy on MySQL dbs and never had a problem with performance,
but when I started writing an app that connected to Microsoft SQL Server
2014 via pyodbc I ran into a serious performance problem.
I'm running the app on my laptop and the database server IS in a VPN but my
internet is f
10 matches
Mail list logo
