Search Results

Search found 27118 results on 1085 pages for 'mysql python'.

Page 28/1085 | < Previous Page | 24 25 26 27 28 29 30 31 32 33 34 35  | Next Page >

  • Corrupted mysql table, cause crash in mysql.h (c++)

    - by Francesco
    i've created a very simple mysql class in c+, but when happen that mysql crash , indexes of tables become corrupted, and all my c++ programs crash too because seems that are unable to recognize corrupted table and allowing me to handle the issue .. Q_RES = mysql_real_query(MY_mysql, tmp_query.c_str(), (unsigned int) tmp_query.size()); if (Q_RES != 0) { if (Q_RES == CR_COMMANDS_OUT_OF_SYNC) cout << "errorquery : CR_COMMANDS_OUT_OF_SYNC " << endl; if (Q_RES == CR_SERVER_GONE_ERROR) cout << "errorquery : CR_SERVER_GONE_ERROR " << endl; if (Q_RES == CR_SERVER_LOST) cout << "errorquery : CR_SERVER_LOST " << endl; LAST_ERROR = mysql_error(MY_mysql); if (n_retrycount < n_retry_limit) { // RETRY! n_retrycount++; sleep(1); cout << "SLEEP - query retry! " << endl; ping(); return select_sql(tmp_query); } return false; } MY_result = mysql_store_result(MY_mysql); B_stored_results = true; cout << "b8" << endl; LAST_affected_rows = (mysql_num_rows(MY_result) + 1); // coult return -1 cout << "b8-1" << endl; the program terminate with a "segmentation fault" after doing the "b8" and before the "b8-1" , Q_RES have no issue even if the table is corrupted.. i would like to know if there is a way to recognize that the table have problems and so then i can run a mysql repair or mysql check .. thanks, Francesco

    Read the article

  • MySQL-python 1.2.3 and OS X 10.5: 64- or 32-bit?

    - by Dave Everitt
    I've been happily using Django and MySQL in development on an existing machine running OS X 10.4 Tiger, and have set up a similar environment in 10.5 Leopard on a new 64-bit MacBook, with a working MySQL and Python 2.6.4. However, now I want them to communicate, easy_install MySQL-python gave ld warnings that the file is not of the required architecture, which led me to test my Python 2.4.6 install (from the Mac OS X disc image): >>> import sys >>> sys.maxint 2147483647 Ah. So my Python install appears to be 32-bit and (I think?) won't install MySQL-python for my 64-bit MySQL. There are lots of hacks out there for MySQL-python on OS X (mostly 1.2.2), but - after hours of reading - I'm pretty sure they won't fix this architecture mismatch. So I'm stuck because I can't decide whether to: give up, remove the 64-bit MySQL install (thorough methods, please?) and use the 32-bit MySQL disc image instead; re-install Python in 64-bit mode from the tarball, --with-universal archs-64-bit and --enable-universalsdk= as detailed in Python.org's 2.6 news. So my questions for anyone who has encountered this issue are: Is installing 64-bit Python on OS X 10.5 worth bothering with? If so, (naive, lazy question!) how are the two required arguments combined? If I just skip along in 32-bit (as on my working setup) what am I missing? I'm after a hassle-free install that's easy to reproduce on other machines (possible student use) so I'd really welcome your opinions, please!

    Read the article

  • http request to cgi python script successful, but the script doesn't seem to run

    - by chipChocolate.py
    I have configured cgi scripts for my apache2 web server. Here is what I want to do: Client uploads the image to the server. (this already works) On success, I want to execute the python script to resize the image. I tried the following and the success function does execute but my python script does not seem to execute: Javascript code that sends the request: var input = document.getElementById('imageLoader'); imageName = input.value; var file = input.files[0]; if(file != undefined){ formData= new FormData(); console.log(formData.length); if(!!file.type.match(/image.*/)){ formData.append("image", file); $.ajax({ url: "upload.php", type: "POST", processData: false, contentType: false, success: function() { var input = document.getElementById('imageLoader'); imageName = input.value; var file = input.files[0]; formData = new FormData(); formData.append("filename", file); $.ajax({ url: "http://localhost/Main/cgi-bin/resize.py", type: "POST", data: formData, processData: false, contentType: false, success: function(data) { console.log(data); } }); // code continues... resize.py: #!/usr/bin/python import cgi import cgitb import Image cgitb.enable() data = cgi.FieldStorage() filename = data.getvalue("filename") im = Image.open("../JS/upload/" + filename) (width, height) = im.size maxWidth = 600 maxHeight = 400 if width > maxWidth: d = float(width) / maxWidth height = int(height / d) width = maxWidth if height > maxHeight: d = float(height) / maxHeight width = int(width / d) height = maxHeight size = (width, height) im = im.resize(size, Image.ANTIALIAS) im.save("../JS/upload/" + filename, quality=100) This is the apache2.conf: <Directory /var/www/html/Main/cgi-bin> AllowOverride None Options +ExecCGI SetHandler cgi-script AddHandler cgi-script .py .cgi Order allow,deny Allow from all </Directory> cgi-bin and python script file permissions: drwxrwxr-x 2 mou mou 4096 Aug 24 03:28 cgi-bin -rwxrwxrwx 1 mou mou 1673 Aug 24 03:28 resize.py Edit: Executing this code $.ajax({ url: "http://localhost/Main/cgi-bin/resize.py", type: "POST", data: formData, // formData = {"filename" : "the filename which was saved in a variable whie the image was uploaded"} processData: false, contentType: false, success: function(data) { alert(data); } }); it alerts the following: <body bgcolor="#f0f0f8"><font color="#f0f0f8" size="-5"> --> <body bgcolor="#f0f0f8"><font color="#f0f0f8" size="-5"> --> --> </font> </font> </font> </script> </object> </blockquote> </pre> </table> </table> </table> </table> </table> </font> </font> </font><body bgcolor="#f0f0f8"> <table width="100%" cellspacing=0 cellpadding=2 border=0 summary="heading"> <tr bgcolor="#6622aa"> <td valign=bottom>&nbsp;<br> <font color="#ffffff" face="helvetica, arial">&nbsp;<br><big><big><strong>&lt;type 'exceptions.TypeError'&gt;</strong></big></big></font></td ><td align=right valign=bottom ><font color="#ffffff" face="helvetica, arial">Python 2.7.6: /usr/bin/python<br>Sun Aug 24 17:24:15 2014</font></td></tr></table> <p>A problem occurred in a Python script. Here is the sequence of function calls leading up to the error, in the order they occurred.</p> <table width="100%" cellspacing=0 cellpadding=0 border=0> <tr><td bgcolor="#d8bbff"><big>&nbsp;</big><a href="file:///var/www/html/Main/cgi-bin/resize.py">/var/www/html/Main/cgi-bin/resize.py</a> in <strong><module></strong>()</td></tr> <tr><td><font color="#909090"><tt>&nbsp;&nbsp;<small>&nbsp;&nbsp;&nbsp;10</small>&nbsp;<br> </tt></font></td></tr> <tr><td><font color="#909090"><tt>&nbsp;&nbsp;<small>&nbsp;&nbsp;&nbsp;11</small>&nbsp;filename&nbsp;=&nbsp;data.getvalue("filename")<br> </tt></font></td></tr> <tr><td bgcolor="#ffccee"><tt>=&gt;<small>&nbsp;&nbsp;&nbsp;12</small>&nbsp;im&nbsp;=&nbsp;Image.open("../JS/upload/"&nbsp;+&nbsp;filename)<br> </tt></td></tr> <tr><td><font color="#909090"><tt>&nbsp;&nbsp;<small>&nbsp;&nbsp;&nbsp;13</small>&nbsp;<br> </tt></font></td></tr> <tr><td><font color="#909090"><tt>&nbsp;&nbsp;<small>&nbsp;&nbsp;&nbsp;14</small>&nbsp;(width,&nbsp;height)&nbsp;=&nbsp;im.size<br> </tt></font></td></tr> <tr><td><small><font color="#909090">im <em>undefined</em>, <strong>Image</strong>&nbsp;= &lt;module 'Image' from '/usr/lib/python2.7/dist-packages/PILcompat/Image.pyc'&gt;, Image.<strong>open</strong>&nbsp;= &lt;function open&gt;, <strong>filename</strong>&nbsp;= '<font color="#c040c0">\xff\xd8\xff\xe0\x00\x10</font>JFIF<font color="#c040c0">\x00\x01\x01\x00\x00\x01\x00\x01\x00\x00\xff\xdb\x00</font>C<font color="#c040c0">\x00\x06\x04\x05\x06\x05\x04\x06\x06\x05\x06\x07\x07\x06\x08\n\x10\n\n\t\t\n\x14\x0e</font>...<font color="#c040c0">\x94\r\x17\x11</font>b<font color="#c040c0">\xcd\xdc\x1a\xfe\xf1\x05\x1b\x15\xd1</font>R<font color="#c040c0">\xce\xe9</font>*<font color="#c040c0">\xb5\x8e</font>b<font color="#c040c0">\x97\x82\x87</font>R<font color="#c040c0">\xf4\xaa</font>K<font color="#c040c0">\x83</font>6<font color="#c040c0">\xbf\xfb</font>0<font color="#c040c0">\xa0\xb6</font>8<font color="#c040c0">\xa9</font>C<font color="#c040c0">\x86\x8d\x96</font>n+E<font color="#c040c0">\xd3\x7f\x99\xff\xd9</font>'</font></small></td></tr></table> <table width="100%" cellspacing=0 cellpadding=0 border=0> <tr><td bgcolor="#d8bbff"><big>&nbsp;</big><a href="file:///usr/lib/python2.7/dist-packages/PIL/Image.py">/usr/lib/python2.7/dist-packages/PIL/Image.py</a> in <strong>open</strong>(fp='../JS/upload/<font color="#c040c0">\xff\xd8\xff\xe0\x00\x10</font>JFIF<font color="#c040c0">\x00\x01\x01\x00\x00\x01\x00\x01\x00\x00\xff\xdb\x00</font>C<font color="#c040c0">\x00\x06\x04\x05\x06\x05\x04\x06\x06\x05\x06</font>...<font color="#c040c0">\x94\r\x17\x11</font>b<font color="#c040c0">\xcd\xdc\x1a\xfe\xf1\x05\x1b\x15\xd1</font>R<font color="#c040c0">\xce\xe9</font>*<font color="#c040c0">\xb5\x8e</font>b<font color="#c040c0">\x97\x82\x87</font>R<font color="#c040c0">\xf4\xaa</font>K<font color="#c040c0">\x83</font>6<font color="#c040c0">\xbf\xfb</font>0<font color="#c040c0">\xa0\xb6</font>8<font color="#c040c0">\xa9</font>C<font color="#c040c0">\x86\x8d\x96</font>n+E<font color="#c040c0">\xd3\x7f\x99\xff\xd9</font>', mode='r')</td></tr> <tr><td><font color="#909090"><tt>&nbsp;&nbsp;<small>&nbsp;1994</small>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;if&nbsp;isPath(fp):<br> </tt></font></td></tr> <tr><td><font color="#909090"><tt>&nbsp;&nbsp;<small>&nbsp;1995</small>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;filename&nbsp;=&nbsp;fp<br> </tt></font></td></tr> <tr><td bgcolor="#ffccee"><tt>=&gt;<small>&nbsp;1996</small>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;fp&nbsp;=&nbsp;builtins.open(fp,&nbsp;"rb")<br> </tt></td></tr> <tr><td><font color="#909090"><tt>&nbsp;&nbsp;<small>&nbsp;1997</small>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;else:<br> </tt></font></td></tr> <tr><td><font color="#909090"><tt>&nbsp;&nbsp;<small>&nbsp;1998</small>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;filename&nbsp;=&nbsp;""<br> </tt></font></td></tr> <tr><td><small><font color="#909090"><strong>fp</strong>&nbsp;= '../JS/upload/<font color="#c040c0">\xff\xd8\xff\xe0\x00\x10</font>JFIF<font color="#c040c0">\x00\x01\x01\x00\x00\x01\x00\x01\x00\x00\xff\xdb\x00</font>C<font color="#c040c0">\x00\x06\x04\x05\x06\x05\x04\x06\x06\x05\x06</font>...<font color="#c040c0">\x94\r\x17\x11</font>b<font color="#c040c0">\xcd\xdc\x1a\xfe\xf1\x05\x1b\x15\xd1</font>R<font color="#c040c0">\xce\xe9</font>*<font color="#c040c0">\xb5\x8e</font>b<font color="#c040c0">\x97\x82\x87</font>R<font color="#c040c0">\xf4\xaa</font>K<font color="#c040c0">\x83</font>6<font color="#c040c0">\xbf\xfb</font>0<font color="#c040c0">\xa0\xb6</font>8<font color="#c040c0">\xa9</font>C<font color="#c040c0">\x86\x8d\x96</font>n+E<font color="#c040c0">\xd3\x7f\x99\xff\xd9</font>', <em>global</em> <strong>builtins</strong>&nbsp;= &lt;module '__builtin__' (built-in)&gt;, builtins.<strong>open</strong>&nbsp;= &lt;built-in function open&gt;</font></small></td></tr></table><p><strong>&lt;type 'exceptions.TypeError'&gt;</strong>: file() argument 1 must be encoded string without NULL bytes, not str <br><tt><small>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</small>&nbsp;</tt>args&nbsp;= ('file() argument 1 must be encoded string without NULL bytes, not str',) <br><tt><small>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</small>&nbsp;</tt>message&nbsp;= 'file() argument 1 must be encoded string without NULL bytes, not str' <!-- The above is a description of an error in a Python program, formatted for a Web browser because the 'cgitb' module was enabled. In case you are not reading this in a Web browser, here is the original traceback: Traceback (most recent call last): File "/var/www/html/Main/cgi-bin/resize.py", line 12, in &lt;module&gt; im = Image.open("../JS/upload/" + filename) File "/usr/lib/python2.7/dist-packages/PIL/Image.py", line 1996, in open fp = builtins.open(fp, "rb") TypeError: file() argument 1 must be encoded string without NULL bytes, not str --> Does this mean that the formData I am sending over is empty?

    Read the article

  • How to make Python Extensions for Windows for absolute beginners

    - by JR
    Hello: I've been looking around the internet trying to find a good step by step guide to extend Python in Windows, and I haven't been able to find something for my skill level. let's say you have some c code that looks like this: #include <stdio.h> #include <math.h> double valuex(float value, double rate, double timex) { float value; double rate, timex; return value / double pow ((1 + rate), (timex)); } and you want to turn that into a Python 3 module for use on a windows (64bit if that makes a difference) system. How would you go about doing that? I've looked up SWIG and Pyrex and in both circumstances they seem geared towards the unix user. With Pyrex I am not sure if it works with Python 3. I'm just trying to learn the basics of programing, using some practical examples. Lastly, if there is a good book that someone can recommend for learning to extend, I would greatly appreciate it. Thank you.

    Read the article

  • Can't get MySQL source query to work using Python mysqldb module

    - by Chris
    I have the following lines of code: sql = "source C:\\My Dropbox\\workspace\\projects\\hosted_inv\\create_site_db.sql" cursor.execute (sql) When I execute my program, I get the following error: Error 1064: You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'source C:\My Dropbox\workspace\projects\hosted_inv\create_site_db.sql' at line 1 Now I can copy and past the following into mysql as a query: source C:\\My Dropbox\\workspace\\projects\\hosted_inv\\create_site_db.sql And it works perfect. When I check the query log for the query executed by my script, it shows that my query was the following: source C:\\My Dropbox\\workspace\\projects\\hosted_inv\\create_site_db.sql However, when I manually paste it in and execute, the entire create_site_db.sql gets expanded in the query log and it shows all the sql queries in that file. Am I missing something here on how mysqldb does queries? Am I running into a limitation. My goal is to run a sql script to create the schema structure, but I don't want to have to call mysql in a shell process to source the sql file. Any thoughts? Thanks!

    Read the article

  • strange syntax error in python, version 2.6 and 3.1

    - by flow
    this may not be an earth-shattering deficiency of python, but i still wonder about the rationale behind the following behavior: when i run source = """ print( 'helo' ) if __name__ == '__main__': print( 'yeah!' ) #""" print( compile( source, '<whatever>', 'exec' ) ) i get :: File "<whatever>", line 6 # ^ SyntaxError: invalid syntax i can avoid this exception by (1) deleting the trailing #; (2) deleting or outcommenting the if __name__ == '__main__':\n print( 'yeah!' ) lines; (3) add a newline to very end of the source. moreover, if i have the source end without a trailing newline right behind the print( 'yeah!' ), the source will also compile without error. i could also reproduce this behavior with python 2.6, so it’s not new to the 3k series. i find this error to be highly irritating, all the more since when i put above source inside a file and execute it directly or have it imported, no error will occur—which is the expected behavior. a # (hash) outside a string literal should always represent the start of a (possibly empty) comment in a python source; moreover, the presence or absence of a if __name__ == '__main__' clause should not change the interpretation of a soure on a syntactical level. can anyone reproduce the above problem, and/or comment on the phenomenon? cheers

    Read the article

  • MySQL table data transformation -- how can I dis-aggregate MySQL time data?

    - by lighthouse65
    We are coding for a MySQL data warehousing application that stores descriptive data (User ID, Work ID, Machine ID, Start and End Time columns in the first table below) associated with time and production quantity data (Output and Time columns in the first table below) upon which aggregate (SUM, COUNT, AVG) functions are applied. We now wish to dis-aggregate time data for another type of analysis. Our current data table design: +---------+---------+------------+---------------------+---------------------+--------+------+ | User ID | Work ID | Machine ID | Event Start Time | Event End Time | Output | Time | +---------+---------+------------+---------------------+---------------------+--------+------+ | 080025 | ABC123 | M01 | 2008-01-24 16:19:15 | 2008-01-24 16:34:45 | 2120 | 930 | +---------+---------+------------+---------------------+---------------------+--------+------+ Reprocessing dis-aggregation that we would like to do would be to transform table content based on a granularity of minutes, rather than the current production event ("Event Start Time" and "Event End Time") granularity. The resulting reprocessing of existing table rows would look like: +---------+---------+------------+---------------------+--------+ | User ID | Work ID | Machine ID | Production Minute | Output | +---------+---------+------------+---------------------+--------+ | 080025 | ABC123 | M01 | 2010-01-24 16:19 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:20 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:21 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:22 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:23 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:24 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:25 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:26 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:27 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:28 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:29 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:30 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:31 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:22 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:33 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:34 | 133 | +---------+---------+------------+---------------------+--------+ So the reprocessing would take an existing row of data created at the granularity of production event and modify the granularity to minutes, eliminating redundant (Event End Time, Time) columns while doing so. It assumes a constant rate of production and divides output by the difference in minutes plus one to populate the new table's Output column. I know this can be done in code...but can it be done entirely in a MySQL insert statement (or otherwise entirely in MySQL)? I am thinking of a INSERT ... INTO construction but keep getting stuck. An additional complexity is that there are hundreds of machines to include in the operation so there will be multiple rows (one for each machine) for each minute of the day. Any ideas would be much appreciated. Thanks.

    Read the article

  • why does b'(and sometimes b' ') show up when I split some HTML source[Python]

    - by Oliver
    I'm fairly new to Python and programming in general. I have done a few tutorials and am about 2/3 through a pretty good book. That being said I've been trying to get more comfortable with Python and proggramming by just trying things in the std lib out. that being said I have recently run into a wierd quirk that I'm sure is the result of my own incorrect or un-"pythonic" use of the urllib module(with Python 3.2.2) import urllib.request HTML_source = urllib.request.urlopen(www.somelink.com).read() print(HTML_source) when this bit is run through the active interpreter it returns the HTML source of somelink, however it prefixes it with b' for example b'<HTML>\r\n<HEAD> (etc). . . . if I split the string into a list by whitespace it prefixes every item with the b' I'm not really trying to accomplish something specific just trying to familiarize myself with the std lib. I would like to know why this b' is getting prefixed also bonus -- Is there a better way to get HTML source WITHOUT using a third party module. I know all that jazz about not reinventing the wheel and what not but I'm trying to learn by "building my own tools" Thanks in Advance!

    Read the article

  • random data using php & mysql

    - by Prakash
    I have mysql database structure like below: CREATE TABLE test ( id int(11) NOT NULL auto_increment, title text NULL, tags text NULL, PRIMARY KEY (id) ); data on field tags is stored as a comma separated text like html,php,mysql,website,html etc... now I need create an array that contains around 50 randomly selected tags from random records. currently I am using rand() to select 15 random mysql data from database and then holding all the tags from 15 records in an array. Then I am using array_rand() for randomizing the array and selecting only 50 random records. $query=mysql_query("select * from test order by id asc, RAND() limit 15"); $tags=""; while ($eachData=mysql_fetch_array($query)) { $additionalTags=$eachData['tags']; if ($tags=="") { $tags.=$additionalTags; } else { $tags.=$tags.",".$additionalTags; } } $tags=explode(",", $tags); $newTags=array(); foreach ($tags as $tag) { $tag=trim($tag); if ($tag!="") { if (!in_array($tag, $newTags)) { $newTags[]=$tag; } } } $random_newTags=array_rand($newTags, 50); Now I have huge records on the database, and because of that; rand() is performing very slow and sometimes it doesn't work. So can anyone let me know how to handle this situation correctly so that my page will work normally.

    Read the article

  • MySQL table data transformation -- how can I dis-aggreate MySQL time data?

    - by lighthouse65
    We are coding for a MySQL data warehousing application that stores descriptive data (User ID, Work ID, Machine ID, Start and End Time columns in the first table below) associated with time and production quantity data (Output and Time columns in the first table below) upon which aggregate (SUM, COUNT, AVG) functions are applied. We now wish to dis-aggregate time data for another type of analysis. Our current data table design: +---------+---------+------------+---------------------+---------------------+--------+------+ | User ID | Work ID | Machine ID | Event Start Time | Event End Time | Output | Time | +---------+---------+------------+---------------------+---------------------+--------+------+ | 080025 | ABC123 | M01 | 2008-01-24 16:19:15 | 2008-01-24 16:34:45 | 2120 | 930 | +---------+---------+------------+---------------------+---------------------+--------+------+ Reprocessing dis-aggregation that we would like to do would be to transform table content based on a granularity of minutes, rather than the current production event ("Event Start Time" and "Event End Time") granularity. The resulting reprocessing of existing table rows would look like: +---------+---------+------------+---------------------+--------+ | User ID | Work ID | Machine ID | Production Minute | Output | +---------+---------+------------+---------------------+--------+ | 080025 | ABC123 | M01 | 2010-01-24 16:19 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:20 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:21 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:22 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:23 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:24 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:25 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:26 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:27 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:28 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:29 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:30 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:31 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:22 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:33 | 133 | | 080025 | ABC123 | M01 | 2010-01-24 16:34 | 133 | +---------+---------+------------+---------------------+--------+ So the reprocessing would take an existing row of data created at the granularity of production event and modify the granularity to minutes, eliminating redundant (Event End Time, Time) columns while doing so. It assumes a constant rate of production and divides output by the difference in minutes plus one to populate the new table's Output column. I know this can be done in code...but can it be done entirely in a MySQL insert statement (or otherwise entirely in MySQL)? I am thinking of a INSERT ... INTO construction but keep getting stuck. An additional complexity is that there are hundreds of machines to include in the operation so there will be multiple rows (one for each machine) for each minute of the day. Any ideas would be much appreciated. Thanks.

    Read the article

  • Error when feeding a mysql db with a python-parsed data

    - by Barnabe
    I use this bit of code to feed some data i have parsed from a web page to a mysql database c=db.cursor() c.executemany( """INSERT INTO data (SID, Time, Value1, Level1, Value2, Level2, Value3, Level3, Value4, Level4, Value5, Level5, ObsDate) VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s, %s)""", clean_data ) The parsed data looks like this (there are several hundred such lines) clean_data = [(161,00:00:00,8.19,1,4.46,4,7.87,4,6.54,null,4.45,6,2010-04-12),(162,00:00:00,7.55,1,9.52,1,1.90,1,4.76,null,0.14,1,2010-04-12),(164,00:00:00,8.01,1,8.09,1,0,null,8.49,null,0.20,2,2010-04-12),(166,00:00:00,8.30,1,4.77,4,10.99,5,9.11,null,0.36,2,2010-04-12)] if i hard code the data as above mySQL accepts my request (except for some quibbles about formatting) but if the variable clean_data is instead defined as the result of the parsing code, like this: cleaner = [(""" $!!'""", ')]'),(' $!!', ') etc etc] def processThis(str,lst): for find, replace in lst: str = str.replace(find, replace) return str clean_data = processThis(data,cleaner) then i get the dreaded "TypeError: not enough arguments for format string" After playing with formatting options for a few hours (I am very new to this) I am confused... what is the difference between the hard coded data and the result of the processThis function as fas as mySQL is concerned? Any idea greatly appreciated...

    Read the article

  • How to create Python module distribution to gracefully fall-back to pure Python code

    - by Craig McQueen
    I have written a Python module, and I have two versions: a pure Python implementation and a C extension. I've written the __init__.py file so that it tries to import the C extension, and if that fails, it imports the pure Python code (is that reasonable?). Now, I'd like to know what is the best way to distribute this module (e.g. write setup.py) so it can be easily used by people with or without the facility to build, or use, the C extension. My experience is limited but I see two possible cases: User does not have MS Visual Studio, or the GCC compiler suite, installed on their machine, to build the C extension User is running IronPython, Jython, or anything other than CPython. I only have used CPython. So I'm not sure how I could distribute this module so that it would work smoothly and be easy to install on those platforms, if they're unable to use the C extension.

    Read the article

  • Python 2 vs Python 3 and Tutorial

    - by MR-J
    Hey guys. I am 12 years old and I have had a small amount of experience with BASIC. I am thinking about learning Python, but I’m not sure if I should learn the 2.6 version or the 3.0 version. I don’t really care about the support for libraries or anything along those lines quite yet. I was wondering if it is easier to code in 3.0 than 2.6. And is it more fun and productive? I would also appreciate it if you could point me in the right direction for a simple yet complete tutorial that is easy to understand and possible teaches Object Oriented Programing. I don’t really care if it teaches OOP or not. One more thing; if I do learn python, is it possible to easily compile a python source code file into a 'stand-alone' .exe file for Windows? I really liked that functionality in BASIC. Thanks!!!!

    Read the article

  • how to build a index table(python dict like) in python with sqlite3

    - by Registered User KC
    Suppose I have one string list may have duplicated items: A B C A A C D E F F I want to make a list can assign an unique index for each item, looks like: 1 A 2 B 3 C 4 D 5 E 6 F now I created sqlite3 database with below SQL statement: CREATE TABLE aa ( myid INTEGER PRIMARY KEY AUTOINCREMENT, name STRING, UNIQUE (myid) ON CONFLICT FAIL, UNIQUE (name) ON CONFLICT FAIL); The plan is insert each row into the database in python. My question is how to handle the error when conflict do happened when insert in python module sqlite3? For example: the program will printing a warning message which item is conflicted and continue next insert action when inserting in python? Thanks

    Read the article

  • MySQL, An Ideal Choice for The Cloud

    - by Bertrand Matthelié
    As the world's most popular web database, MySQL has quickly become the leading database for the cloud, with most providers offering MySQL-based services. Normal 0 false false false EN-US X-NONE X-NONE /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-qformat:yes; mso-style-parent:""; mso-padding-alt:0cm 5.4pt 0cm 5.4pt; mso-para-margin:0cm; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:11.0pt; font-family:"Calibri","sans-serif"; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-fareast-font-family:"Times New Roman"; mso-fareast-theme-font:minor-fareast; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin; mso-bidi-font-family:"Times New Roman"; mso-bidi-theme-font:minor-bidi;} Access our Resource Kit to discover: Why MySQL has become the leading database in the cloud, and how it addresses the critical attributes of cloud-based deployments How ISVs rely on MySQL to power their SaaS offerings Best practices to deploy the world’s most popular open source database in public and private clouds Normal 0 false false false EN-US X-NONE X-NONE You will also find out how you can leverage MySQL together with Hadoop and other technologies to unlock the value of Big Data, either on-premise or in the cloud. Access white papers, webinars, case studies and other resources in /* Style Definitions */ table.MsoNormalTable {mso-style-name:"Table Normal"; mso-tstyle-rowband-size:0; mso-tstyle-colband-size:0; mso-style-noshow:yes; mso-style-priority:99; mso-style-qformat:yes; mso-style-parent:""; mso-padding-alt:0cm 5.4pt 0cm 5.4pt; mso-para-margin:0cm; mso-para-margin-bottom:.0001pt; mso-pagination:widow-orphan; font-size:11.0pt; font-family:"Calibri","sans-serif"; mso-ascii-font-family:Calibri; mso-ascii-theme-font:minor-latin; mso-fareast-font-family:"Times New Roman"; mso-fareast-theme-font:minor-fareast; mso-hansi-font-family:Calibri; mso-hansi-theme-font:minor-latin; mso-bidi-font-family:"Times New Roman"; mso-bidi-theme-font:minor-bidi;} our Resource Kit now!

    Read the article

  • How to proceed setting up a secondary mysql linux slave?

    - by Algorist
    I have a mysql database master and slave in production. I want to setup additional mysql slave. There is around 15 Terabyte of data in the database and there are MYISAM and InnoDB tables in the database. I am thinking of below options: Shutdown master database and copy the mysql data folder to secondary slave. Can Innodb tables be copied like this? Run flush table with read lock, scp the file to new slave and unlock the table and this is possible for myisam tables, can I do the same for innodb tables too? Thanks for looking at the question.

    Read the article

  • How to generate the right password format for Apache2 authentication in use with DBD and MySQL 5.1?

    - by Walkman
    I want to authenticate users for a folder from a MySQL 5.1 database with AuthType Basic. The passwords are stored in plain text (they are not really passwords, so doesn't matter). The password format for apache however only allows for SHA1, MD5 on Linux systems as described here. How could I generate the right format with an SQL query ? Seems like apache format is a binary format with a lenght of 20, but the mysql SHA1 function return 40 long. My SQL query is something like this: SELECT CONCAT('{SHA}', BASE64_ENCODE(SHA1(access_key))) FROM user_access_keys INNER JOIN users ON user_access_keys.user_id = users.id WHERE name = %s where base64_encode is a stored function (Mysql 5.1 doesn't have TO_BASE64 yet). This query returns a 61 byte BLOB which is not the same format that apache uses. How could I generate the same format ? You can suggest other method for this too. The point is that I want to authenticate users from a MySQL5.1 database using plain text as password.

    Read the article

  • What could cause sudden crash of a MySQL 5.0.67 installation?

    - by Alex R
    I have an old Ubuntu 8.10 32-bit with MySQL 5.0.67. There's 5.7GB of data in it and it grows by about 100MB every day. About 3 days ago, the MySQL instance begin dying suddenly and quitely (no log entry) during the nightly mysqldump. What could be causing it? Upgrading MySQL is a long-term project for me, unless there happens to be a specific bug in 5.0.67 then I guess I'll just need to reprioritize. I'm hoping somebody might be familiar with this problem since this is a fairly popular version bundled with Ubuntu 8.10. Thanks

    Read the article

  • Python script is exiting with no output and I have no idea why

    - by Adam Tuttle
    I'm attempting to debug a Subversion post-commit hook that calls some python scripts. What I've been able to determine so far is that when I run post-commit.bat manually (I've created a wrapper for it to make it easier) everything succeeds, but when SVN runs it one particular step doesn't work. We're using CollabNet SVNServe, which I know from the documentation removes all environment variables. This had caused some problems earlier, but shouldn't be an issue now. Before Subversion calls a hook script, it removes all variables - including $PATH on Unix, and %PATH% on Windows - from the environment. Therefore, your script can only run another program if you spell out that program's absolute name. The relevant portion of post-commit.bat is: echo -------------------------- >> c:\svn-repos\company\hooks\svn2ftp.out.log set SITENAME=staging set SVNPATH=branches/staging/wwwroot/ "C:\Python3\python.exe" C:\svn-repos\company\hooks\svn2ftp.py ^ --svnUser="svnusername" ^ --svnPass="svnpassword" ^ --ftp-user=ftpuser ^ --ftp-password=ftppassword ^ --ftp-remote-dir=/ ^ --access-url=svn://10.0.100.6/company ^ --status-file="C:\svn-repos\company\hooks\svn2ftp-%SITENAME%.dat" ^ --project-directory=%SVNPATH% "staging.company.com" %1 %2 >> c:\svn-repos\company\hooks\svn2ftp.out.log echo -------------------------- >> c:\svn-repos\company\hooks\svn2ftp.out.log When I run post-commit.bat manually, for example: post-commit c:\svn-repos\company 12345, I see output like the following in svn2ftp.out.log: -------------------------- args1: c:\svn-repos\company args0: staging.company.com abspath: c:\svn-repos\company project_dir: branches/staging/wwwroot/ local_repos_path: c:\svn-repos\company getting youngest revision... done, up-to-date -------------------------- However, when I commit something to the repo and it runs automatically, the output is: -------------------------- -------------------------- svn2ftp.py is a bit long, so I apologize but here goes. I'll have some notes/disclaimers about its contents below it. #!/usr/bin/env python """Usage: svn2ftp.py [OPTION...] FTP-HOST REPOS-PATH Upload to FTP-HOST changes committed to the Subversion repository at REPOS-PATH. Uses svn diff --summarize to only propagate the changed files Options: -?, --help Show this help message. -u, --ftp-user=USER The username for the FTP server. Default: 'anonymous' -p, --ftp-password=P The password for the FTP server. Default: '@' -P, --ftp-port=X Port number for the FTP server. Default: 21 -r, --ftp-remote-dir=DIR The remote directory that is expected to resemble the repository project directory -a, --access-url=URL This is the URL that should be used when trying to SVN export files so that they can be uploaded to the FTP server -s, --status-file=PATH Required. This script needs to store the last successful revision that was transferred to the server. PATH is the location of this file. -d, --project-directory=DIR If the project you are interested in sending to the FTP server is not under the root of the repository (/), set this parameter. Example: -d 'project1/trunk/' This should NOT start with a '/'. 2008.5.2 CKS Fixed possible Windows-related bug with tempfile, where the script didn't have permission to write to the tempfile. Replaced this with a open()-created file created in the CWD. 2008.5.13 CKS Added error logging. Added exception for file-not-found errors when deleting files. 2008.5.14 CKS Change file open to 'rb' mode, to prevent Python's universal newline support from stripping CR characters, causing later comparisons between FTP and SVN to report changes. """ try: import sys, os import logging logging.basicConfig( level=logging.DEBUG, format='%(asctime)s %(levelname)s %(message)s', filename='svn2ftp.debug.log', filemode='a' ) console = logging.StreamHandler() console.setLevel(logging.ERROR) logging.getLogger('').addHandler(console) import getopt, tempfile, smtplib, traceback, subprocess from io import StringIO import pysvn import ftplib import inspect except Exception as e: logging.error(e) #capture the location of the error frame = inspect.currentframe() stack_trace = traceback.format_stack(frame) logging.debug(stack_trace) print(stack_trace) #end capture sys.exit(1) #defaults host = "" user = "anonymous" password = "@" port = 21 repo_path = "" local_repos_path = "" status_file = "" project_directory = "" remote_base_directory = "" toAddrs = "[email protected]" youngest_revision = "" def email(toAddrs, message, subject, fromAddr='[email protected]'): headers = "From: %s\r\nTo: %s\r\nSubject: %s\r\n\r\n" % (fromAddr, toAddrs, subject) message = headers + message logging.info('sending email to %s...' % toAddrs) server = smtplib.SMTP('smtp.company.com') server.set_debuglevel(1) server.sendmail(fromAddr, toAddrs, message) server.quit() logging.info('email sent') def captureErrorMessage(e): sout = StringIO() traceback.print_exc(file=sout) errorMessage = '\n'+('*'*80)+('\n%s'%e)+('\n%s\n'%sout.getvalue())+('*'*80) return errorMessage def usage_and_exit(errmsg): """Print a usage message, plus an ERRMSG (if provided), then exit. If ERRMSG is provided, the usage message is printed to stderr and the script exits with a non-zero error code. Otherwise, the usage message goes to stdout, and the script exits with a zero errorcode.""" if errmsg is None: stream = sys.stdout else: stream = sys.stderr print(__doc__, file=stream) if errmsg: print("\nError: %s" % (errmsg), file=stream) sys.exit(2) sys.exit(0) def read_args(): global host global user global password global port global repo_path global local_repos_path global status_file global project_directory global remote_base_directory global youngest_revision try: opts, args = getopt.gnu_getopt(sys.argv[1:], "?u:p:P:r:a:s:d:SU:SP:", ["help", "ftp-user=", "ftp-password=", "ftp-port=", "ftp-remote-dir=", "access-url=", "status-file=", "project-directory=", "svnUser=", "svnPass=" ]) except getopt.GetoptError as msg: usage_and_exit(msg) for opt, arg in opts: if opt in ("-?", "--help"): usage_and_exit() elif opt in ("-u", "--ftp-user"): user = arg elif opt in ("-p", "--ftp-password"): password = arg elif opt in ("-SU", "--svnUser"): svnUser = arg elif opt in ("-SP", "--svnPass"): svnPass = arg elif opt in ("-P", "--ftp-port"): try: port = int(arg) except ValueError as msg: usage_and_exit("Invalid value '%s' for --ftp-port." % (arg)) if port < 1 or port > 65535: usage_and_exit("Value for --ftp-port must be a positive integer less than 65536.") elif opt in ("-r", "--ftp-remote-dir"): remote_base_directory = arg elif opt in ("-a", "--access-url"): repo_path = arg elif opt in ("-s", "--status-file"): status_file = os.path.abspath(arg) elif opt in ("-d", "--project-directory"): project_directory = arg if len(args) != 3: print(str(args)) usage_and_exit("host and/or local_repos_path not specified (" + len(args) + ")") host = args[0] print("args1: " + args[1]) print("args0: " + args[0]) print("abspath: " + os.path.abspath(args[1])) local_repos_path = os.path.abspath(args[1]) print('project_dir:',project_directory) youngest_revision = int(args[2]) if status_file == "" : usage_and_exit("No status file specified") def main(): global host global user global password global port global repo_path global local_repos_path global status_file global project_directory global remote_base_directory global youngest_revision read_args() #repository,fs_ptr #get youngest revision print("local_repos_path: " + local_repos_path) print('getting youngest revision...') #youngest_revision = fs.youngest_rev(fs_ptr) assert youngest_revision, "Unable to lookup youngest revision." last_sent_revision = get_last_revision() if youngest_revision == last_sent_revision: # no need to continue. we should be up to date. print('done, up-to-date') return if last_sent_revision or youngest_revision < 10: # Only compare revisions if the DAT file contains a valid # revision number. Otherwise we risk waiting forever while # we parse and uploading every revision in the repo in the case # where a repository is retroactively configured to sync with ftp. pysvn_client = pysvn.Client() pysvn_client.callback_get_login = get_login rev1 = pysvn.Revision(pysvn.opt_revision_kind.number, last_sent_revision) rev2 = pysvn.Revision(pysvn.opt_revision_kind.number, youngest_revision) summary = pysvn_client.diff_summarize(repo_path, rev1, repo_path, rev2, True, False) print('summary len:',len(summary)) if len(summary) > 0 : print('connecting to %s...' % host) ftp = FTPClient(host, user, password) print('connected to %s' % host) ftp.base_path = remote_base_directory print('set remote base directory to %s' % remote_base_directory) #iterate through all the differences between revisions for change in summary : #determine whether the path of the change is relevant to the path that is being sent, and modify the path as appropriate. print('change path:',change.path) ftp_relative_path = apply_basedir(change.path) print('ftp rel path:',ftp_relative_path) #only try to sync path if the path is in our project_directory if ftp_relative_path != "" : is_file = (change.node_kind == pysvn.node_kind.file) if str(change.summarize_kind) == "delete" : print("deleting: " + ftp_relative_path) try: ftp.delete_path("/" + ftp_relative_path, is_file) except ftplib.error_perm as e: if 'cannot find the' in str(e) or 'not found' in str(e): # Log, but otherwise ignore path-not-found errors # when deleting, since it's not a disaster if the file # we want to delete is already gone. logging.error(captureErrorMessage(e)) else: raise elif str(change.summarize_kind) == "added" or str(change.summarize_kind) == "modified" : local_file = "" if is_file : local_file = svn_export_temp(pysvn_client, repo_path, rev2, change.path) print("uploading file: " + ftp_relative_path) ftp.upload_path("/" + ftp_relative_path, is_file, local_file) if is_file : os.remove(local_file) elif str(change.summarize_kind) == "normal" : print("skipping 'normal' element: " + ftp_relative_path) else : raise str("Unknown change summarize kind: " + str(change.summarize_kind) + ", path: " + ftp_relative_path) ftp.close() #write back the last revision that was synced print("writing last revision: " + str(youngest_revision)) set_last_revision(youngest_revision) # todo: undo def get_login(a,b,c,d): #arguments don't matter, we're always going to return the same thing try: return True, "svnUsername", "svnPassword", True except Exception as e: logging.error(e) #capture the location of the error frame = inspect.currentframe() stack_trace = traceback.format_stack(frame) logging.debug(stack_trace) #end capture sys.exit(1) #functions for persisting the last successfully synced revision def get_last_revision(): if os.path.isfile(status_file) : f=open(status_file, 'r') line = f.readline() f.close() try: i = int(line) except ValueError: i = 0 else: i = 0 f = open(status_file, 'w') f.write(str(i)) f.close() return i def set_last_revision(rev) : f = open(status_file, 'w') f.write(str(rev)) f.close() #augmented ftp client class that can work off a base directory class FTPClient(ftplib.FTP) : def __init__(self, host, username, password) : self.base_path = "" self.current_path = "" ftplib.FTP.__init__(self, host, username, password) def cwd(self, path) : debug_path = path if self.current_path == "" : self.current_path = self.pwd() print("pwd: " + self.current_path) if not os.path.isabs(path) : debug_path = self.base_path + "<" + path path = os.path.join(self.current_path, path) elif self.base_path != "" : debug_path = self.base_path + ">" + path.lstrip("/") path = os.path.join(self.base_path, path.lstrip("/")) path = os.path.normpath(path) #by this point the path should be absolute. if path != self.current_path : print("change from " + self.current_path + " to " + debug_path) ftplib.FTP.cwd(self, path) self.current_path = path else : print("staying put : " + self.current_path) def cd_or_create(self, path) : assert os.path.isabs(path), "absolute path expected (" + path + ")" try: self.cwd(path) except ftplib.error_perm as e: for folder in path.split('/'): if folder == "" : self.cwd("/") continue try: self.cwd(folder) except: print("mkd: (" + path + "):" + folder) self.mkd(folder) self.cwd(folder) def upload_path(self, path, is_file, local_path) : if is_file: (path, filename) = os.path.split(path) self.cd_or_create(path) # Use read-binary to avoid universal newline support from stripping CR characters. f = open(local_path, 'rb') self.storbinary("STOR " + filename, f) f.close() else: self.cd_or_create(path) def delete_path(self, path, is_file) : (path, filename) = os.path.split(path) print("trying to delete: " + path + ", " + filename) self.cwd(path) try: if is_file : self.delete(filename) else: self.delete_path_recursive(filename) except ftplib.error_perm as e: if 'The system cannot find the' in str(e) or '550 File not found' in str(e): # Log, but otherwise ignore path-not-found errors # when deleting, since it's not a disaster if the file # we want to delete is already gone. logging.error(captureErrorMessage(e)) else: raise def delete_path_recursive(self, path): if path == "/" : raise "WARNING: trying to delete '/'!" for node in self.nlst(path) : if node == path : #it's a file. delete and return self.delete(path) return if node != "." and node != ".." : self.delete_path_recursive(os.path.join(path, node)) try: self.rmd(path) except ftplib.error_perm as msg : sys.stderr.write("Error deleting directory " + os.path.join(self.current_path, path) + " : " + str(msg)) # apply the project_directory setting def apply_basedir(path) : #remove any leading stuff (in this case, "trunk/") and decide whether file should be propagated if not path.startswith(project_directory) : return "" return path.replace(project_directory, "", 1) def svn_export_temp(pysvn_client, base_path, rev, path) : # Causes access denied error. Couldn't deduce Windows-perm issue. # It's possible Python isn't garbage-collecting the open file-handle in time for pysvn to re-open it. # Regardless, just generating a simple filename seems to work. #(fd, dest_path) = tempfile.mkstemp() dest_path = tmpName = '%s.tmp' % __file__ exportPath = os.path.join(base_path, path).replace('\\','/') print('exporting %s to %s' % (exportPath, dest_path)) pysvn_client.export( exportPath, dest_path, force=False, revision=rev, native_eol=None, ignore_externals=False, recurse=True, peg_revision=rev ) return dest_path if __name__ == "__main__": logging.info('svnftp.start') try: main() logging.info('svnftp.done') except Exception as e: # capture the location of the error for debug purposes frame = inspect.currentframe() stack_trace = traceback.format_stack(frame) logging.debug(stack_trace[:-1]) print(stack_trace) # end capture error_text = '\nFATAL EXCEPTION!!!\n'+captureErrorMessage(e) subject = "ALERT: SVN2FTP Error" message = """An Error occurred while trying to FTP an SVN commit. repo_path = %(repo_path)s\n local_repos_path = %(local_repos_path)s\n project_directory = %(project_directory)s\n remote_base_directory = %(remote_base_directory)s\n error_text = %(error_text)s """ % globals() email(toAddrs, message, subject) logging.error(e) Notes/Disclaimers: I have basically no python training so I'm learning as I go and spending lots of time reading docs to figure stuff out. The body of get_login is in a try block because I was getting strange errors saying there was an unhandled exception in callback_get_login. Never figured out why, but it seems fine now. Let sleeping dogs lie, right? The username and password for get_login are currently hard-coded (but correct) just to eliminate variables and try to change as little as possible at once. (I added the svnuser and svnpass arguments to the existing argument parsing.) So that's where I am. I can't figure out why on earth it's not printing anything into svn2ftp.out.log. If you're wondering, the output for one of these failed attempts in svn2ftp.debug.log is: 2012-09-06 15:18:12,496 INFO svnftp.start 2012-09-06 15:18:12,496 INFO svnftp.done And it's no different on a successful run. So there's nothing useful being logged. I'm lost. I've gone way down the rabbit hole on this one, and don't know where to go from here. Any ideas?

    Read the article

  • 100,000 complex structures that are accessed frequently by 100,000 users

    - by Saad
    If you are required to store 100,000 complex structures that are accessed frequently by 100,000 users, which of the following solutions would you use and why? Memcached, In-code python objects, Redis, or a relational database (MySQL). With the little knowledge that I have I think that memcached and In-code python object will not store permanent persistent data. so they don't qualify as the right answer for such a problem. And for complex data structures its best to use Redis. Please correct me if I am wrong.

    Read the article

  • Closest to “Mathematica Graphics[]" drawing environment for Python

    - by 500
    Being only familiar with Mathematica and its Graphics, I have now to learn to draw Graphics using Python for a server. Mostly conditional combination of simple shape. What would be a package for Python that make drawing Graphics as close as possible as the Mathematica Graphics environment ? For Example, I would need to do such thing as in : http://mathematica.stackexchange.com/questions/1010/2d-gaussian-distribution-of-squares-coordinates#comment2475_1010

    Read the article

  • Open Source Survey: Oracle Products on Top

    - by trond-arne.undheim
    Oracle continues to work with the open source community to bring the most innovative and productive software to market (more). Oracle products received the most votes in several key categories of the 2010 Linux Journal Reader's Choice Awards. With over 12,000 technologists reporting, these product earned top spots: Best Office Suite: OpenOffice.org Best Single Office Program: OpenOffice.org Writer Best Database: MySQL Best Virtualization Solution: VirtualBox "As the leading open source technology and service provider, Oracle continues to work with the community stakeholders to rapidly innovate many open source products for use in fully tested production environments," says Edward Screven, Oracle's chief corporate architect. "Supporting open source is important to Oracle and our customers, and we continue to invest in it." According to a recent report by the Linux Foundation, Oracle is one of the top ten contributors to the Linux Kernel. Oracle also contributes millions of lines of code to these important projects: OpenJDK: 7,002,579 Eclipse: 1,800,000 (#3 in active committers) MySQL: 5,073,113 NetBeans: 7,870,446 JSF: 701,980 Apache MyFaces Trinidad: 1,316,840 Hudson: 1,209,779 OpenOffice.org: 7,500,000

    Read the article

  • Pitfalls of using MySQL as your database choice?

    - by Sergio
    I've read online on multiple occassions that MySQL is a bad database. The places I've read this include some threads on Reddit, but they never seem to delve in on why it's a poor product. Is there any truth to this claim? I've never used it beyond a very simple CRUD scenario, and that was for a university project during my second year. What pitfalls, if any, are there when choosing MySQL as your database?

    Read the article

  • Python Forgiveness vs. Permission and Duck Typing

    - by darkfeline
    In Python, I often hear that it is better to "beg forgiveness" (exception catching) instead of "ask permission" (type/condition checking). In regards to enforcing duck typing in Python, is this try: x = foo.bar except AttributeError: pass else: do(x) better or worse than if hasattr(foo, "bar"): do(foo.bar) else: pass in terms of performance, readability, "pythonic", or some other important factor?

    Read the article

< Previous Page | 24 25 26 27 28 29 30 31 32 33 34 35  | Next Page >