规范空白
来自 PostgreSQL 维基
跳转到导航跳转到搜索
此 SQL 函数规范字符串中的空格,移除任何前导或尾随空格,并将任何内部空白缩减为每次出现一个空格字符。它在创建域时很有用。
CREATE OR REPLACE FUNCTION normalize_space(TEXT)
RETURNS TEXT
IMMUTABLE
LANGUAGE SQL
AS $$
SELECT regexp_replace(
trim($1),
E'\\s+',
' ',
'g'
);
$$;
CREATE OR REPLACE FUNCTION normalize_space(text)
RETURNS text LANGUAGE plperl AS $$
$_ = $_[0];
s/(^\s+|\s+$)//g;
s/\s+/ /og;
return $_;
$$;
因此,如果您想更新/插入到表中并规范之前该操作的文本,您可以在每次插入之前创建一个触发器,如下所示
CREATE TABLE pepito (wordy text, moody text, bar text);
DROP IF EXISTS trigger n_space on pepito;
CREATE TRIGGER n_space BEFORE INSERT OR UPDATE ON pepito FOR EACH ROW EXECUTE PROCEDURE n_space();
此函数触发器查找文本类型的字段,并使用替换表达式修改 NEW 值。
CREATE OR REPLACE FUNCTION n_space()
RETURNS trigger
VOLATILE
LANGUAGE plperl
AS $$
my $rs = spi_exec_query(<<" EOT");
SELECT
a.attname as name,
format_type(a.atttypid,pg_attribute.atttypmod)
FROM
pg_class AS c
JOIN
pg_attribute AS a
ON (
a.attrelid = c.oid
AND c.oid = '$_TD->{table_name}'::regclass
AND a.attstorage <> 'p'
AND format_type(a.atttypid,pg_attribute.atttypmod) = 'text'
EOT
my $rs_rows = $rs->{processed}-1;
foreach my $rn (0 .. $rs_rows){
my $row = $rs->{rows}[$rn]->{name};
$_TD->{new}{$row} =~ s/(^\s+|\s+$)//g;
$_TD->{new}{$row} =~ s/\s+/ /g;
}
return "MODIFY";
$$;
TRUNCATE TABLE pepito;
INSERT INTO pepito
VALUES
('asd asd AS asd asf ',' asd 4t45gr g er ergt',' asd sa'),
('asd asd AS asdbhfg',' asd 4t45gr g er ergt',' asd sa'),
(' asd asd AS asd asf ',' asd 4t45gr g er ergt',' asd sa');
SELECT * FROM pepito;