oralcle解析xml求助

[zzx129728xxx]

数据库中A表中T_A指标存储的内容为：
"<?xml version="1.0" encoding="GBK"?>
<records columns="BAUAR`BAUAA`BAUAS`BAUAB`BAUAQ`BAUAC`BAUAD`BAUAE`BAUAF">
<record I9999="1" state="" edit="1">AA001`2019.01.01``0103`010304`0106`驾驶员`1`</record>
<record I9999="2" state="" edit="1">AA002`2019.01.01``0103`010315`0104`党务`1`</record>
</records>"
怎么使用 sql数据查询解析出下列数据：

[nayi_224]

这种数据可以直接截取字符串

create or replace function f_str_start_end(f_str varchar2, f_start varchar2, f_end varchar2)
  return varchar2 as
  i varchar2(2000);
begin
  if f_str is null then
    return null;
  end if;
  i := substr(f_str,
              instr(f_str, f_start) + length(f_start),
              instr(f_str, f_end, instr(f_str, f_start)) -
              instr(f_str, f_start) - length(f_start));
  return i;
exception
  when others then
    return '-1';
end;

with tab1 as (
select '"<?xml version="1.0" encoding="GBK"?>
<records columns="BAUAR`BAUAA`BAUAS`BAUAB`BAUAQ`BAUAC`BAUAD`BAUAE`BAUAF">
  <record I9999="1" state="" edit="1">AA001`2019.01.01``0103`010304`0106`驾驶员`1`</record>
  <record I9999="2" state="" edit="1">AA002`2019.01.01``0103`010315`0104`党务`1`</record>
</records>"' aa from dual
)
select f_str_start_end(t1.aa, '<record I9999="1" state="" edit="1">', '</record>') from tab1 t1;

剩下的列可以用regexp_substr慢慢拼