This section is about how to tell analog the format of your logfile. Most
people don't need to do this because analog can detect the format
automatically -- try it first and see, because you will save yourself a
lot of trouble! But if you do need to specify the log format explicitly, here
is how to do it.
The basic command to specify a log format looks like
LOGFORMAT format
-- we'll discuss what the formats can be in a minute. Or if you are using the
Apache server, you will probably find it more convenient to use
APACHELOGFORMAT apacheformat
instead.
The LOGFORMAT and APACHELOGFORMAT
commands only apply to logfiles specified with a LOGFILE command
later in the same configuration file. So you must put the
LOGFORMAT above the LOGFILE to which it refers. This
way, different logfiles can have different formats, like this:
In this example, log1 is in format1, log2 and
log3 are in format2, and log0 isn't in either
format -- analog will try and detect which format it's in.
The APACHELOGFORMAT command is followed by the
LogFormat from your Apache httpd.conf file. For example,
if your httpd.conf contained the following lines:
(Use parentheses instead of quotes round the argument if the argument already
contains quotes.) Analog
understands all Apache log formats, with the exception that it won't parse
Apache's "%...{format}t" construction for customised
times: if you have this construction, you will have to use ordinary
LOGFORMAT instead.
The possible formats for use with the
LOGFORMAT command are of two
types. First there are some symbolic words, and then there are log format
strings. We'll look at the words first.
There are format words for all the built-in formats
analog knows about.
You might need one of these words if your logfile is in a standard format, but
analog can't detect which format it's in for some reason; for example, maybe
the first line is corrupt; or maybe analog can't tell whether you're using
North American or international dates. So for example
LOGFORMAT COMMON
will select common format; you can also have COMBINED,
REFERRER, BROWSER, EXTENDED,
MICROSOFT-NA (North American date format),
MICROSOFT-INT (international date format),
WEBSITE-NA, WEBSITE-INT,
MS-EXTENDED (Microsoft's attempt at extended format),
WEBSTAR-EXTENDED (WebSTAR's version of extended format),
MS-COMMON (a buggy version of common format in some versions
of Microsoft software), NETSCAPE, WEBSTAR or
MACHTTP. All these
formats were defined at the end of the previous
section. You can also use the special word AUTO to return to
automatic detection.
If your logfile is not in one of the recognised
formats, you can tell analog
about your format using a log format string. You only ever need this if your
logfile has lines which are not in one of the standard formats. (And even if it
isn't in a standard format, if you're using the Apache web server, you will
find APACHELOGFORMAT easier.)
The format string consists of a template for the logfile line, with the
various fields and special characters replaced by codes as follows. Please
note that these codes are case sensitive -- for example, %b is
completely different from %B!
%S
host (the client hostname, or address of the computer
making the request)
%s
numerical IP address of client (if recorded in a
separate field; used when %S is empty)
%r
file requested
%q
query string (part of filename after ?, if
recorded in a separate field)
%B
browser
%A
browser with +'s instead of spaces
%f
referrer
%u
user (tip: a cookie can usefully be defined as
%u too)
%v
virtual host (the server hostname, also called the
virtual domain)
%d
day of the month
%m
month in digits
%M
month, three letter English abbreviation
%y
year, last two digits
%Y
year, four digits
%Z
year, two or four digits (less efficient)
%h
hour of the day
%n
minute of the hour
%a
a or A for am, or p or
P for pm, if %h is in the 12-hour clock. (So to
match "am" you need %am and to match "AM"
you need %aM)
%U
"Unix time" (seconds since beginning of 1970,
GMT). If it includes decimals, use %U.%j
%b
number of bytes transferred
%t
processing time in seconds
%T
processing time in milliseconds
%c
HTTP status code
%C
code words used instead of HTTP status code in some
servers -- only used internally
%j
junk: ignore this field (field can be empty too)
%w
white space: spaces or tabs
%W
optional white space
%%
% sign
\n
new line
\t
tab stop
\\
single backslash
So for example, the common log format, which looks like
jay.bird.com - fred [25/Dec/1998:17:45:35 +0000]
"GET /~sret1/ HTTP/1.0" 200 1243
(except all on one line)
could be represented by the LOGFORMAT command
In other words, it's just the sample line but with the hostname replaced by
%S, the username by %u etc.
(The parentheses are needed because the argument contains spaces.)
Or take another example: if you had lines which looked like
Remember: if you have trouble writing a LOGFORMAT string, you can
turn debugging on, and analog will
report where each line was corrupt. If you still have trouble, you can write
to the analog-help mailing list.
A logfile can sometimes have lines in several different formats. So you can
specify several LOGFORMAT commands in a row, and they will all
apply to the next logfile. This is also useful if the format of your logfile
changes half way through. So in this example:
log1 has lines in both common and combined format, whereas
log2 and log3 have lines just in the format in the
previous example.
If you specify several formats, analog tries to match each line to the first
format first, then if that fails the next, and so on, so the order of the
formats is important. Usually you want to specify the most common one
first, to minimise the time spent trying to match lines to inappropriate
formats.
I suggested above that any logfile which
doesn't have a LOGFORMAT
command earlier in the same configuration file is auto-detected. But this
isn't quite true. Actually such logfiles get a special format called the
default log format. The default format starts off as auto-detection,
but you can change it if you want with the DEFAULTLOGFORMAT
command. This command works exactly the same as the LOGFORMAT
command -- it understands the same formats, and if you have several
DEFAULTLOGFORMAT commands, they accumulate in the same way. The
difference is that they don't need to be put in any particular place. (There
is also APACHEDEFAULTLOGFORMAT, which has the same effect but uses
the Apache LogFormat strings.)
Here log0 actually gets the default log format. If there are no
DEFAULTLOGFORMAT commands, the default will be auto-detection. But
if there are DEFAULTLOGFORMAT commands, even in another
configuration file, that will be the format of log0.
The times you need to use the DEFAULTLOGFORMAT instead of the
LOGFORMAT are if you want to change the format of logfiles which
aren't given in a LOGFILE command -- for example, ones specified on
the command line, or dragged onto the program icon on a Mac, or compiled in.
It is also useful to use the DEFAULTLOGFORMAT if your logfiles are
always in the same format, so that you don't have to worry about putting in
enough LOGFORMATs in the right places.
The "Unix time", %U, is always recorded in GMT. So you
will probably need to use a
LOGTIMEOFFSET
command to convert to your local timezone. Also, it's just the integer part of
the time, so if you have decimals you will have to use %U.%j .
The log formats which analog can handle are those which are known as
instantaneously decipherable: in practice, this means that the character
which terminates a string can never occur in the string. So for example, in
common format, which looks like
if the hostname ever contained a space, the line would be marked as corrupt,
because analog terminates the host at the first space, not at the
first occurrence of space-dash-space, and then the rest of the line wouldn't
match. Of course, hostnames should never contain spaces, so this shouldn't be a
problem. There are a couple of other restrictions: if there is any date or
time information, then the year, month, date, hour and minute must all be
present: and the same information may not occur twice in the format (so you
can't have both %m and %M, for example, because these
both represent the month; make one of them a %j to have it ignored).
Sometimes you need to read one of the fields in a
logfile, but not analyse it.
For example, if you have a separate common log and referrer log, the referrer
log might look like
But the requests for /~sret1/analog/ would already have been
counted when reading the main logfile, so you don't want to count them again
now. You get round this by specifying a * in that item in the
format string, like this:
LOGFORMAT (%f -> %*r)
A tip: sometimes it is more efficient to specify two or more adjacent fields
to ignore with a single %j, as long as the whole group ends with a
recognisable character. So common format is more efficiently specified as
-- in the date and time [25/Dec/1998:17:45:35 +0000], the seconds
and the timezone can be ignored with a single %j, extending until
the close-bracket.
Another tip: %j can also be used to ignore whole lines, rather than
just fields analog doesn't use. For example, the extended log format ignores
lines beginning with # by using
LOGFORMAT #%j
and the Microsoft format ignores lines corresponding to FTP requests with
LOGFORMAT (%*S, %*u, %m/%d/%y, %h:%n:%j, %j)
If those formats had not been used, the lines would have been incorrectly
marked as corrupt.
Finally, both for reference and as examples, here is
a list of all the fixed formats that analog understands, together with the
example lines from the previous section
and their built-in definitions (split over two lines where necessary).
12/25/98 17:45:35 OK jay.bird.com /~sret1/ 1243
LOGFORMAT (%m/%d/%y\t%h:%n:%j \t%C%w%S\t%r\t%b)
The extended log, Netscape log and WebSTAR log don't have any built-in
formats: analog constructs their formats from their header lines.
Go to the analog home page.