# Round Decimal

From MidrangeWiki

## Summary

The following are the RPG/LE fully free-form definitions and instructions needed for using the **Round Decimal** service procedure. This service procedure simply allows the caller to round a decimal number to a specified number of decimals. Optionally, custom rounding is also available—such as always rounding down or always rounding up.

By Dave Clark

## Service Prototype

Place the following in a separate copybook for inclusion in both the caller and the service program source members.

**free //============================================================================== // This procedure rounds the passed number to the specified number of decimal // positions as the returned result. Maximum result decimals is nine and that // is also the default. For normal rounding (where .4 or below rounds down // while .5 or above rounds up) do not pass the third parm. Optionally, pass // the maximum decimal value to round down. Examples: // .9 always rounds down, .0 always rounds up, and .4 is normal rounding. //============================================================================== dcl-pr GenUtl_roundDecimal packed(31:9); pNumber packed(33:11) const; pDecimals uns(3) const options(*nopass:*omit); pRoundAt packed(1:1) const options(*nopass); end-pr;

## Service Procedure

Place the following in a service program source member.

**free ctl-opt NoMain AlwNull(*UsrCtl) Debug Option(*SrcStmt:*NoDebugIo) DatFmt(*ISO) TimFmt(*ISO); //============================================================================== // This procedure rounds the passed number to the specified number of decimal // positions as the returned result. Maximum result decimals is nine and that // is also the default. For normal rounding (where .4 or below rounds down // while .5 or above rounds up) do not pass the third parm. Optionally, pass // the maximum decimal value to round down. Examples: // .9 always rounds down, .0 always rounds up, and .4 is normal rounding. //============================================================================== dcl-proc GenUtl_roundDecimal export; dcl-pi *n packed(31:9); pNumber packed(33:11) const; pDecimals uns(3) const options(*nopass:*omit); pRoundAt packed(1:1) const options(*nopass); end-pi; dcl-s adjustBy int(10:0); dcl-s iRoundBy packed(1:1) inz(.5); if %parms() < %parmnum(pDecimals) // if parm not passed or %addr(pDecimals) = *null // or it was omitted or pDecimals > 9; // or is greater than internal limit adjustBy = 10 ** 9; // default adjust by internal limit else; // else adjustBy = 10 ** pDecimals; // adjust by decimal positions specified endif; if %parms() < %parmnum(pRoundAt) // if parm not passed or %addr(pRoundAt) = *null; // or it was omitted else; // use default rounding, else iRoundBy = .9 - %abs(pRoundAt); // calculate the rounding factor endif; return (%int(pNumber * adjustBy + iRoundBy) / adjustBy); // return rounded val end-proc;