Question

我为Postgres编写了以下函数，当通过pgAdmin安装时，它可以在我的本地Windows机器上正常工作。当我尝试将它添加到我的基于Linux的服务器安装时，它会抛出编译错误：

QUERY：SELECT $ 1（split_part（$ 2，''，$ 3））语境：PL / PgSQL函数中的SQL语句“第23行附近的分裂词”

CREATE OR REPLACE FUNCTION splitwords(text, int)
  RETURNS text AS
$BODY$
DECLARE
inwords ALIAS FOR $1;
posn INTEGER;
existcount INTEGER;
incurrdataid ALIAS FOR $2;
currdataid INTEGER;
currwordid INTEGER;
length INTEGER;
wordpos INTEGER;
newword TEXT;
BEGIN
currdataid:=incurrdataid;
currdataid:=currdataid-1; --corrects for auto-increment error
posn:=1;
WHILE posn<11 LOOP
IF split_part(inwords,' ',posn)='' THEN
-- If no more words are available
    EXIT;
ELSE
--If not at the end of the words
    IF (SELECT wordID FROM words WHERE word=split_part(inwords,' ',posn))>0 THEN
    --If word is already in lexicon
        currwordid:=(SELECT wordID FROM words WHERE word=split_part(inwords,' ',posn))::INTEGER;
        existcount:= (SELECT count FROM words WHERE word=split_part(inwords,' ',posn))::INTEGER;
        UPDATE words SET count=existcount+1 WHERE word=split_part(inwords,' ',posn);
        INSERT INTO wordsdata(wordid,dataid) VALUES (currwordid,currdataid);
        posn:=posn+1;
    ELSE
    --If word is new
        newword=split_part(inwords,' ',posn);
        INSERT INTO words(word,count) VALUES (newword,1);
        currwordid:=(SELECT wordID FROM words WHERE word=split_part(inwords,' ',posn))::INTEGER;
        INSERT INTO wordsdata(wordid,dataid) VALUES (currwordid,currdataid);
        length:=length(split_part(inwords,' ',posn));
        wordpos:=1;
        WHILE wordpos<(length+1) LOOP
            INSERT INTO searchchar(searchstr,wordid) VALUES (substring(split_part(inwords,' ',posn),1,wordpos),currwordid);
            wordpos=wordpos+1;
        END LOOP;
        posn:=posn+1;
    END IF;
END IF;
END LOOP;

RETURN 'rows added';
END
$BODY$
  LANGUAGE plpgsql VOLATILE
  COST 100;

有谷歌和文档的拖网，但我找不到任何相关的东西。我很迷茫！您可以提供的任何帮助都非常感谢！

Answer 1

没有读取您的函数，但错误消息通常意味着变量与列的名称相同。使用_前缀变量名通常是避免消息的好方法，它使plpgsql函数更具可读性。

此外，代码中至少出现一次=，看起来应该是:=。 plpgsql对这些东西很宽容，但你不应该依赖它。

PL / PgSQL混淆错误

1 个答案: