Get the decimal part of a number (positive and negative safe)

Why x - INT(x) gives the wrong cents on negative refunds

For positive 14.75, both 14.75 - INT(14.75) and 14.75 - TRUNC(14.75) return 0.75. On a negative refund of -14.75, INT floors downward to -15, so -14.75 - (-15) returns +0.25 instead of 75 cents!

TRUNC chops toward zero (-14), so ROUND(A2 - TRUNC(A2), 6) returns -0.75 (signed fraction) and ABS(ROUND(A2 - TRUNC(A2), 6)) returns 0.75 (unsigned fractional-unit magnitude).

Unsigned and signed decimal extraction formulas

Wrap in ROUND(..., 6) to strip binary floating-point dust (e.g., 0.750000000000002).

=ABS(ROUND(A2 - TRUNC(A2), 6))

Signed fractional remainder

Preserves the minus sign on negative amounts so TRUNC(A2) + SignedFraction reconstructs A2.

=ROUND(A2 - TRUNC(A2), 6)

Inputs and interpretation

For a dollar amount, 0.75 is a fractional dollar, not the numeric count 75 cents. Multiply by 100 and use an explicit currency rounding rule to obtain cents. Rounding the remainder to six places can turn 0.9999999 into 1, so reconstruction is approximate at that precision. The floor-based nonnegative fraction is useful for other purposes; it is only unsuitable when the intended output is a signed toward-zero remainder.

Further reading

• Microsoft Support — TRUNC function