ERROR
The attribute key cannot be found when processing: Table: 'dbo_viu_isb', Column: 'city', Value: ' '. The attribute is 'City'.
OR
The attribute key cannot be found when processing: Table: 'dbo_viu_isb', Column: 'customername', Value: 'ﾊﾟﾅｿﾆｯｸ（株）'. The attribute is 'Customername'.
CAUSE
The problem is caused by the Unicode nature of the SPACE character in the text field in the database. In the field are for example Kanji characters and unicode has about 18 characters for SPACE:
http://www.fileformat.info/info/unicode/category/Zs/list.htm
There is no way to tell Analysis Services or SQL Server the difference between these spaces. In short: collation, case sensitivity, accent sensitivity, width sensitivity or Kana sensitivity have no influence.
WORKAROUND
Convert all the kind of SPACE characters to one SPACE character, e.g:
replace(ad_.ad_name,' ',' ') as ad_name
This appears to do nothing, but the first empty space finds ALL space characters, e.g. ASCII 32 and Unicode 12288 and the second parameter converts it to ONE space (Unicode 32)