如果从 shell 启动，则 PostgreSQL 上的 QuerySet 上的 Django UnicodeEncodeError 会额外出现

发布于 2024-11-04 21:44:51 字数 1264 浏览 0 评论 0原文

大家好我在两台机器上使用 Django 1.2.5 和 Python 2.6：Ubuntu 11.04 和 Debian Lenny。第一个是我的本地开发人员，第二个是远程服务器。我使用 django-fts 和 postgresql （数据库是远程的并且对于这两种安装都是唯一的）来进行全文搜索查询。有趣的是我可以在本地机器上使用俄语符号进行 fts 查询。在远程服务器上，它在评估我的额外内容时给了我 UnicodeEncodeError 。

在 django-fts postgre 后端中，我找到了这段代码：

    ts_query = "plainto_tsquery('%s','%s')" % (self.language, unicode(query).replace("'", "''"))
    where = '%s @@ %s' % (qn(self.vector_field.column), ts_query)
    select = {}
    order = []
    if rank_field is not None:
        select[rank_field] = 'ts_rank(%s.%s, %s, %d)' % (qn(self.model._meta.db_table), qn(self.vector_field.column), ts_query, rank_normalization)
        order = ['-%s' % rank_field]
    return qs.extra(select=select, where=[where], order_by=order)`

其中查询包含 u'\u0430' 或 u"а" （西里尔文），因此到目前为止它都以 utf-8 编码。

尝试从 extras 组装查询时，django/db/models/sql/compiler.py 的第 489 行出现错误： result.append('(%s)' % str(col))

所以，我尝试了所有编码/解码的东西。 LANG 变量在两台计算机上均为“ru_RU.UTF-8”，在 settings.py 中为 DEFAULT_CHARSET。没有更多的想法了。有什么帮助吗？

更新：我刚刚发现我的应用程序在两台计算机上的执行方式相同。区别在于从 shell (python manage.py runserver localhost:8000) 执行和从 eclipse 的 pydev 环境启动调试模式。那么，有人可以告诉我关于我的编码问题手动启动和 pydev 调试启动之间是否有区别吗？

TIA。彼得

原文

Hello all
I'm using Django 1.2.5 with Python 2.6 on two machines: Ubuntu 11.04 and Debian Lenny. First one is my local dev and second is remote server. I'm using django-fts and postgresql (database is remote and unique for both installations) to make full text search queries. The intriguing thing is I can do fts query with russian symbols in it on local machine all right. On remote server it gives me UnicodeEncodeError while evaluating my extra.

In django-fts postgre backend I find this code:

    ts_query = "plainto_tsquery('%s','%s')" % (self.language, unicode(query).replace("'", "''"))
    where = '%s @@ %s' % (qn(self.vector_field.column), ts_query)
    select = {}
    order = []
    if rank_field is not None:
        select[rank_field] = 'ts_rank(%s.%s, %s, %d)' % (qn(self.model._meta.db_table), qn(self.vector_field.column), ts_query, rank_normalization)
        order = ['-%s' % rank_field]
    return qs.extra(select=select, where=[where], order_by=order)`

where query contains u'\u0430' or u"а" (cyrillic) so it encoded in utf-8 all right so far.

Error appears in django/db/models/sql/compiler.py on line 489 while trying to assemble a query from extras:
result.append('(%s)' % str(col))

So, I tried all of coding/decoding stuff. LANG variable is "ru_RU.UTF-8" on both computers and in settings.py as DEFAULT_CHARSET. Have no more ideas. Any help?

UPDATE: I've just found out that my application executes the same way on both computers. The difference is between execution from shell (python manage.py runserver localhost:8000) and debug-mode start from eclipse's pydev environment. So, can some one tell me if there's difference between manual start and pydev debug start concerning my encoding problems?

TIA.
Petr

分享到QQ

分享到微博