【SAS 小技巧】用K函數處理中文字串

在Facebook中的「SAS戰術應用精研社」，網友偶會提到使用文字函數處理中文時遇到一些的狀況，如果使用一搬文字函數如compress、scan、substr、index等可能是得到一堆亂碼，或是找不到設定文字，又或是找到非預期中的字，或等，例如鄭姓網友所提的問題：

data a1;
input string $20.;
cards;
洛瓦
北極光
野生動物
黃色小鴨
窈窕曲線包
手藝有限
;
data a2;
set a1;
idx1=index(string,'孕');
idx2=index(string,'嬰');
run;
proc print data=a2;
run;


這6個字串裡面沒有「看到」任何「孕」或「嬰」，但用index函數卻得到前3個字串有孕字，後三個字串有嬰字。這一類的問題主要為中文為雙字元編碼，SAS傳統的文字函數適合處理單字元編碼的函數。

在SAS Base建有處理雙位元的函數，在說明檔案中，SAS將這一系列稱為『K Functions』，詳細K函數請見SAS官方網頁。

最後擷取說明檔中部分K函數的說明：

函數	說明
KCOMPARE	Returns the result of a comparison of character expressions.
KCOMPRESS	Removes specified characters from a character expression.
KCOUNT	Returns the number of double-byte characters in an expression.
KINDEX	Searches a character expression for a string of characters.
KINDEXC	Searches a character expression for specified characters.
KLEFT	Left-aligns a character expression by removing unnecessary leading DBCS blanks and SO/SI.
KLENGTH	Returns the length of an argument.
KLOWCASE	Converts all letters in an argument to lowercase.
KPROPCASE	Converts Chinese, Japanese, Korean, Taiwanese (CJKT) characters.
KPROPCHAR	Converts special characters to normal characters.
KPROPDATA	Removes or converts unprintable characters.
KREVERSE	Reverses a character expression.
KRIGHT	Right-aligns a character expression by trimming trailing DBCS blanks and SO/SI.
KSCAN	Selects a specified word from a character expression.
KSTRCAT	Concatenates two or more character expressions.
KSUBSTR	Extracts a substring from an argument.
KSUBSTRB	Extracts a substring from an argument according to the byte position of the substring in the argument.
KTRANSLATE	Replaces specific characters in a character expression.
KTRIM	Removes trailing DBCS blanks and SO/SI from character expressions.
KTRUNCATE	Truncates a string to a specified length in byte unit without breaking multibyte characters.
KUPCASE	Converts all letters in an argument to uppercase.
KUPDATE	Inserts, deletes, and replaces character value contents.
KUPDATEB	Inserts, deletes, and replaces the contents of the character value according to the byte position of the character value in the argument.
KVERIFY	Returns the position of the first character that is unique to an expression.