Are you using RPG/400 or RPG IV (ILE)? If the latter you can use two asterisks for "raise to the power of": <pre> C Eval Result = 125 ** (1/3) </pre> If you're using RPG/400 then you could create an RPG IV program that accepts the value and root to be calculated and returns the result: <pre> D* Prototype D ThisPgm pr ExtPgm('THISPGM') D iValue 15p 6 D iRoot 1p 0 D oResult 15p 6 D* Procedure Interface (*ENTRY PLIST)... D ThisPgm pi D iValue 15p 6 D iRoot 1p 0 D oResult 15p 6 C* Calculate the root. Note, if iRoot is zero then we close the program down C* otherwise we calculate the root and leave it open... C If iRoot = *Zeros C Eval *InLR = *On C Else C Eval oResult = iValue ** (1 / iRoot) C EndIf C* Return to the calling program or close down... C Return </pre> Another alternative if you don't want to use RPG IV or ILE is to use embedded SQL as in the following example: <pre> C Z-ADD*ZEROS RESULT 156 C Z-ADD125 VALUE 156 C Z-ADD3 ROOT 10 ** C/EXEC SQL C+ SET :RESULT = :VALUE ** (1 / :ROOT) C/END-EXEC ** C RESULT DSPLY ** C MOVE *ON *INLR </pre> Hope it helps Best regards Jonathan ================================================== Note that the system supports exponentiation. The _POWER MI builtin is available: <pre> H Debug H dftactgrp( *NO ) H actgrp( *CALLER ) * Procedure Prototype (this.program) D Exponent PR extpgm( 'EXPONENT' ) * Procedure Interface D Exponent PI D n s 30p 7 D e s 5p 3 D r s 30p 7 * Procedure Prototype (_POWER MI builtin) D power pr 8f extproc( '_POWER' ) D 8f value D 8f value C seton LR C eval n = 100.47 C eval e = 2.0 C eval r = *zero C eval r = power( n : e ) C dump C eval r = power( 11.23 : 1.7 ) C dump C return </pre> That example calls the builtin twice -- once with declared variables and once with constants. The builtin operates on 64-bit floating-point values. I suspect that you can achieve just about any degree of precision you'd like. Tom

Are you using RPG/400 or RPG IV (ILE)? If the latter you can use two asterisks for "raise to the power of": <pre> C Eval Result = 125 ** (1/3) </pre> If you're using RPG/400 then you could create an RPG IV program that accepts the value and root to be calculated and returns the result: <pre> D* Prototype D ThisPgm pr ExtPgm('THISPGM') D iValue 15p 6 D iRoot 1p 0 D oResult 15p 6 D* Procedure Interface (*ENTRY PLIST)... D ThisPgm pi D iValue 15p 6 D iRoot 1p 0 D oResult 15p 6 C* Calculate the root. Note, if iRoot is zero then we close the program down C* otherwise we calculate the root and leave it open... C If iRoot = *Zeros C Eval *InLR = *On C Else C Eval oResult = iValue ** (1 / iRoot) C EndIf C* Return to the calling program or close down... C Return </pre> Another alternative if you don't want to use RPG IV or ILE is to use embedded SQL as in the following example: <pre> C Z-ADD*ZEROS RESULT 156 C Z-ADD125 VALUE 156 C Z-ADD3 ROOT 10 ** C/EXEC SQL C+ SET :RESULT = :VALUE ** (1 / :ROOT) C/END-EXEC ** C RESULT DSPLY ** C MOVE *ON *INLR </pre> Hope it helps Best regards Jonathan ================================================== Note that the system supports exponentiation. The _POWER MI builtin is available: <pre> H Debug H dftactgrp( *NO ) H actgrp( *CALLER ) * Procedure Prototype (this.program) D Exponent PR extpgm( 'EXPONENT' ) * Procedure Interface D Exponent PI D n s 30p 7 D e s 5p 3 D r s 30p 7 * Procedure Prototype (_POWER MI builtin) D power pr 8f extproc( '_POWER' ) D 8f value D 8f value C seton LR C eval n = 100.47 C eval e = 2.0 C eval r = *zero C eval r = power( n : e ) C dump C eval r = power( 11.23 : 1.7 ) C dump C return </pre> That example calls the builtin twice -- once with declared variables and once with constants. The builtin operates on 64-bit floating-point values. I suspect that you can achieve just about any degree of precision you'd like. Tom