odbcQuery , memory.size
On Tue, 18 Mar 2008, Lukas Rohrer wrote:
Dear R cracks I am trying to fetch 38 Tables from 38 ESRI Geodatabases through an ODBC connection.
On Windows, using RODBC?
I stored the 38 channels in a list and the 38 tablenames in an other list.
Why? It looks like you only need the channel once.
With a
for(i in 1:38) .... sqlFetch(....) I try to read the tables into a third
list. But always after a certain amount of rounds (mostly 16) i get
Fehler in odbcQuery(channel, query, rows_at_time) :
Calloc konnte nicht (67108864 von 1) Speicher zuteilen
Zus??zlich: Warning message:
Reached total allocation of 1021Mb: see help(memory.size)
When I keep on trying the rounds will stop at 1.
Any sugestions?
Buy more meomry. 1Gb is very low these days for working with databases. But take a look at what happened when you tried 16, by using object.size() on the objects you created. I suspect they are too large to store in memory in your small computer. If so, you need to store them on disc: there are several ways to do that such as package SQLiteDF or RODBC to a single database. If not, you may succeed by retrieving say 13,13,12 in 3 sessions and saving the workspaces, then loading all three workspaces into one session.
Sincerely Lukas -- -- Arbeit: Tel: +41 44 632 55 98 ETH Z??ich Institut f?? terrestrische ??osysteme Professur Wald??ologie Universit??strasse 22 CH-8092 Z??ich [[alternative HTML version deleted]]
Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595