Well it not make any impression. 1...e5 remains the classical approach and Super GM's when looking for balanced games they prefer this instead of the Sicilian. The Sicilian offers a much more risk and that sometimes in a long round robin tournament remains a suspensory factor to choose as a main weapon.
If you don't use chess tempo data then there is another source to confirm this.
Yeah, that's the one I was using. 
I didn't want to say it's name because chess.com censored chess tempo before.
Do any other databases confirm this?
Are there any other databases that let you search by rating like this?
Please!
I only have a low quality free downloaded database on my computer, updated with the weekly TWIC files.
Without filter: 1222447 games with 1. e4, 526259 Sicilians, 269870 Open Games
With filter 2600+ (both players): 24900 games with 1.e4, 9781 Sicilians, 8816 Open Games.
The chess tempo online database has:
Without filter: 1419530 games with 1.e4, 625169 Sicilians and 302087 Open Games.
With filter 2600+ (both players): 24008 games with 1.e4, 8826 Sicilians and 9531 Open Games.
Looks like both databases deliver roughly the same results.
Look OP..your database is fine and the fact that 2600 GM's prefer 1...e5 from the Sicilian is not a phenomenon! I tried to explain this to you referring to the risk Sicilian games have from the more balanced symmetrical pawn structures. The last years after 2000 when Kramnik brought in to fashion the Berlin Wall ending the percent of games starting with 1.e4 e5 increased dramatically.
I don't understand your insist to this statistic evidence!
I was looking around in a chess database and I noticed that when i set the rating floor at 2600 there were actually more games that responded to 1 e4 with 1 ...e5 than 1 ...c5. Can anybody confirm or deny this? I'm trying to find out if it is a problem with the database i'm using, or a real phenomenon.