Shadow

CJKE Japanese log

Tried to use SQL to combine eye movement data and the SEG info (both word-based and char-based). It turned out to be a big mistake. The following request took 18 hours to run, and the result still has errors -- with the Edison stories, the left(trialID,4) is not enough to distinguish the trials. Other trialID's are like "JA04.seg" or simply "JA04". But I wasn't able to find the function to drop everything after the "." There is the LEFT function, but no RIGHT function.

SELECT data.*, seg.char AS charsegid, seg.x1 AS charx1, seg.x2 AS charx2, seg.y1 AS chary1, seg.y2 AS chary2, seg.letter AS chartext, [seg.preceding sequence] AS prevchar, [seg.following sequence] AS nexthar, [seg.charatype] AS chartype, [seg.letterid] AS charid INTO letterWord
FROM Emxml2dat AS data LEFT JOIN LetterSeg AS seg ON (left(ucase(data.trialid),4)=left(ucase(seg.trialid),4)) AND (data.x-58-data.currentCond between seg.x1 and seg.x2) AND (data.y-62 between seg.y1-40 and seg.y2+40);

posted by Gary Feng @ 4:50 PM