The default first line of a CDX file is :
CDX A b e a m s c k r V v D d g M n

The letters use in dat files and cdx files are as follows :

A canonized url
B news group
C rulespace category ***
D compressed dat file offset
F canonized frame
G multi-columm language description (* soon)
H canonized host
I canonized image
J canonized jump point
K Some weird FBIS what's changed kinda thing
L canonized link
M meta tags (AIF) *
N massaged url
P canonized path
Q language string
R canonized redirect
U uniqness ***
V compressed arc file offset *
X canonized url in other href tages
Y canonized url in other src tags
Z canonized url found in script
a original url **
b date **
c old style checksum *
d uncompressed dat file offset
e IP **
f frame *
g file name
h original host
i image *
j original jump point
k new style checksum *
l link *
m mime type of original document *
n arc document length *
o port
p original path
r redirect *
s response code *
t title *
v uncompressed arc file offset *
x url in other href tages *
y url in other src tags *
z url found in script *
# comment

* in alexa-made dat file
** in alexa-made dat file meta-data line
*** future data