[Info-ingres] precision

[Roy Hann]

Thursday, November 15, 2018, 1:51:01 PM, you wrote:

> Yep worked a treat. Thanks Roy.

> I was getting satisfaction from this but I think it has quite a bit more
> overhead.

Generally tests involving floats should be done in terms of an
acceptable error (sometimes referred to as epsilon). e.g.

... having abs(sum(source_amount))-float4(0.01) < float4(.00001)

Roy Hann
Rational Commerce Ltd.
T +44 20 8691 2089
"Ingres development, tuning and training experts"