Yeah, you may be interested in a total, but I think it's addtotals because you want the sum of each row. On another note, you should consider the way your data is saved to file and indexed. Have a look at the logging best practices to get some basic ideas. The reason I'm saying this is because while you may know what this data is supposed to stand for (and how to interpret it), it is completely unlabeled (except for the host, filename/source/sourcetype), so there is very little in this data you can use later to correlate it with other information you have. In short, you (and anyone else for that matter) will never be able to easily use this data other than to look at the sum of each row here. If you somehow can, you should really try to use key=value pairs, include the actual timestamp of those events in these files (they can be in any format - epoch, human readable, anything really), and at least add some information about what each row stands for. Much of this could be done if you included a header in your file and an identifier for each row, so that it looks more like this: "time","component","temperature_1","value_2",... <any actual timestamp>,rack1,16,17,... <any actual timestamp>,rack2,16,17,... <any actual timestamp>,psu,16,17,... See how these little additions radically improve the amount of information you can retrieve? It will be hard to work with your data (for everyone else even more) if you just index a simple row of undescript numbers. ... View more