On 2010-10-07 14:48, Mark Bucciarelli wrote:

Hi,

I'm getting a "write error" when running a fairly large and complex query (five tables and subselect in where).

I run

$ mclient -l sql -d <db> -u <usr> < t.sql password: ... % 6, 26, 1, 103, 229, 5 # length write error $ echo $? 1

The write error happens when mclient wants to write a character that it got from the server (i.e. part of the data you're printing) that cannot be converted to the output character set. For instance, if you have some Japanese of Greek (or whatever) character in your data, and you're attempting to write to a latin-1 (a.k.a. iso-8859-1) terminal. Try running mclient with the -Eutf-8 option to tell it to "convert" to UTF-8 (i.e. don't convert anything). The output may not look good on your terminal, but you shouldn't get a write error either. (You might want to redirect the output to a file.)

I've run the exact query on SQLite, so I'm pretty sure the SQL is good.

This is Feb2010. I'm thinking about upgrading, but I'm close to a release so I'd prefer not to.

Does this description remind anyone of a commit to Jun2010 that may have fixed such an issue?

Trying to boil the (complex) SQL down to a small reproduceable test case will take a lot of time.

Is there another approach I can use to debug that would be faster?

There is nothing in the merovingian log.

Thanks,

m

---

Beautiful is writing same markup. Internet Explorer 9 supports standards for HTML5, CSS3, SVG 1.1, ECMAScript5, and DOM L2 & L3. Spend less time writing and rewriting code and more time creating great experiences on the web. Be a part of the beta today. http://p.sf.net/sfu/beautyoftheweb _______________________________________________ Monetdb-developers mailing list Monetdb-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-developers