我的解决办法是:

psql -U your_user your_db < your.file.here.sql  2>&1|more

这样我就可以读取错误消息

我希望这能对大家有所帮助。

In my recent experience, it's possible to get this error when the real problem has nothing to do with escape characters or newlines. In my case, I had created a dump from database A with pg_dump -a -t table_name > dump.sql and was trying to restore it to database B with psql < dump.sql (after updating the proper env vars, of course) What I finally figured out was that the dump, though it was data-only (the -a option, so that the table structure isn't explicitly part of the dump), was schema-specific. That meant that without manually modifying the dump, I couldn't use a dump generated from schema1.table_name to populate schema2.table_name. Manually modifying the dump was easy, the schema is specified in the first 15 lines or so.

对我来说，与源数据库不同的是ENCODING和LOCALE。 一旦我放弃了目标DB并重新创建它，它就可以正常工作。

我知道这是一个旧帖子，但我遇到了另一个解决方案:postgis没有安装在我的新版本上，这导致我在pg_dump上出现同样的错误

我过去也遇到过这种错误。Pavel是正确的，这通常是pg_restore创建的脚本中某些东西失败的标志。由于所有的“/N”错误，您在输出的顶部看不到真正的问题。我建议:

插入单个小表(例如，pg_restore . table) ——表full_database =订单。转储>个订单。转储) 如果你没有一个小的，那么从恢复脚本中删除一堆记录-我只是确保./是最后一行被加载(例如，打开订单。转储和删除一堆记录) 观察标准输出，一旦发现问题，就可以随时放弃 表格和重载

在我的情况下，我还没有安装“hstore”扩展，所以脚本在顶部失败。我在目标数据库上安装了hstore，然后就可以继续工作了。

检查表中的列和备份文件中的列是否合适