Date: 2016-02-12 02:22:35 +0100
From: Anthony Damico <>
To: GDK devs <>
Version: 11.21.13 (Jul2015-SP2)
CC: ajdamico, @hannesmuehleisen, @njnes

Last updated: 2016-03-25 09:58:57 +0100

Comment 21782

Date: 2016-02-12 02:22:35 +0100
From: Anthony Damico <>

User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:44.0) Gecko/20100101 Firefox/44.0
Build Identifier:

should the second result be NaN or Inf?

the result of MAX( ) appears to work

if the column contains both -Inf and also Inf, then i believe the column should return NaN

Reproducible: Always

Steps to Reproduce:

library(MonetDBLite)
library(MonetDB.R)
con <- dbConnect( MonetDBLite() , tempdir() )
x <- mtcars
x[1,1] <- Inf
dbWriteTable( con , 'x' , x )

dbGetQuery( con , "SELECT MIN( mpg ) , AVG( mpg ) , MAX( mpg ) FROM x" )

Actual Results:

dbGetQuery( con , "SELECT MIN( mpg ) , AVG( mpg ) , MAX( mpg ) FROM x" )
L1 L2 L3
1 10.4 NaN Inf

Expected Results:

L2 = Inf, i think? but not sure if this is my R bias

thanks!

Comment 21833

Date: 2016-02-26 16:03:14 +0100
From: @sjoerdmullender

This is an instance of a bigger problem: SQL doesn't allow for Inf (+ or -) or NaN. And hence, MonetDB doesn't allow for those values either. The R interface is now a backdoor through which those values can be introduced to the MonetDB internals, which are not prepared to handle them. In other words, garbage in, garbage out.

Comment 21835

Date: 2016-02-26 17:00:15 +0100
From: MonetDB Mercurial Repository <>

Changeset c3894dbe6e00 made by Hannes Muehleisen hannes@cwi.nl in the MonetDB repo, refers to this bug.

For complete details, see http//devmonetdborg/hg/MonetDB?cmd=changeset;node=c3894dbe6e00

Changeset description:

Bug #3922, do not import NaN or Inf

Comment 21952

Date: 2016-03-25 09:58:57 +0100
From: @sjoerdmullender

Jul2015-SP3 has been released.