最后一行代码应是：
 
REPLACE _1dxxf WITH '_1DXXF' FOR NOT '_'$_1dxx AND ('_'$k_1d OR '_'$k_1x OR '_'$k_1t) AND RECNO()=i

没这样用过

[此贴子已经被作者于2020-9-13 15:13编辑过]

RECNO()是t2的当前记录号
RECNO()+1就是说：
t2的当前行与t1相对应行的下一行，t2行1时t1行2, t2行2时t1行3, ......

用 SELECT SQL 语句，测试一下：
SELECT a.jnh, (b._1dxt>0 AND a._1dxt==b._1dxt) _1dxx, a._1dxt;
    FROM t1 a, t1 b;
    WHERE a.jnh==(b.jnh+1)

与
REPLACE ALL t1._1dxx  WITH t2._1dxt>0 AND t1._1dxt==t2._1dxt
  
SELECT jnh,;
       IIF(_1dxx,"_1DXX",IIF(_1dxt>0,"_1DXXF","      ")) _1dxx,;
       ICASE(_1dxt==1,"_1D",_1dxt==2,"_1X",_1dxt==3,"_1T","   ") _1dxt FROM t1
结果一样，代码更简捷!

[此贴子已经被作者于2020-9-13 18:55编辑过]

以上方法代码简捷，运行效率高，我还得一一来学。
按4楼的方法：
表:

CLOSE TABLES all

USE b1 IN 0 ALIAS t1
BLANK FIELDS _1d, _1x,_1t,_2d, _2x,_2t ALL 
USE b1 IN 0 ALIAS t2 AGAIN 

SELECT t1

SET RELATION TO RECNO()+1 INTO "t2"


REPLACE ALL t2._1d  WITH IIF(t2.a1>t1.a1 ,"_1D", "");
            t2._1x  WITH IIF(t2.a1<t1.a1 ,"_1X", "");
            t2._1t  WITH IIF(t2.a1=t1.a1 ,"_1t", "");
            t2._2d  WITH IIF(t2.a2>t1.a2 ,"_2D", "");
            t2._2x  WITH IIF(t2.a2<t1.a2 ,"_2X", "");
            t2._2t  WITH IIF(t2.a2=t1.a2 ,"_2t", "")
           
SET RELATION TO
SELECT * FROM t1
BROWSE

没有问题，且运行比以前快多了
注：
"_1D"-下1行比上1行对应值大
"_1X"-下1行比上1行对应值小
"_1T"-下1行比上1行对应值相同



下面用变量就出问题

CLOSE TABLES all
USE b1 IN 0 ALIAS t1
BLANK FIELDS _1d, _1x,_1t,_2d, _2x,_2t ALL 
USE b1 IN 0 ALIAS t2 AGAIN 

SELECT t1

SET RELATION TO RECNO()+1 INTO "t2"

p1=ALLTRIM("t1.")+ALLTRIM('a')+ALLTRIM('1')
p11=&p1
k1=ALLTRIM("t2.")+ALLTRIM('a')+ALLTRIM('1')
k11=&k1
p2=ALLTRIM("t1.")+ALLTRIM('a')+ALLTRIM('2')
p22=&p2
k2=ALLTRIM("t2.")+ALLTRIM('a')+ALLTRIM('2')
k22=&k2
REPLACE ALL t2._1d  WITH IIF(k11>p11 ,"_1D", "");
            t2._1x  WITH IIF(k11<p11 ,"_1X", "");
            t2._1t  WITH IIF(k11=p11 ,"_1t", "");
            t2._2d  WITH IIF(k22>p22 ,"_2D", "");
            t2._2x  WITH IIF(k22<p22 ,"_2X", "");
            t2._2t  WITH IIF(k22=p22 ,"_2t", "")
           
SET RELATION TO
SELECT * FROM t1
BROWSE

问题在哪？

[此贴子已经被作者于2020-9-13 18:45编辑过]

数据记录多时可测试一下 SET RELATION TO 与 SELECT SQL 的速度，有时表面看语句简单并不就高效。

k11、p11等变量只获取了一条记录的数据
REPLACE ALL 是处理所有记录数据

p1=ALLTRIM("t1.")+ALLTRIM('a')+ALLTRIM('1')
p11=&p1
相当于直接写 p11 = t1.a1

BLANK FIELDS _1d, _1x,_1t,_2d, _2x,_2t ALL
这句可以不要，因下面的 REPLACE ALL 处理了