Code:

nawk '{ K[$1,$2,$3]++ }
$5 ~ /^(14|9[89])$/ && !t[$1,$2,$3,$5]++ {
T[$1,$2,$3]++
R[$1,$2,$3] = $1 FS $2 FS $3 FS $6
} END {
for (k in K)
 if ((K[k] < 6) && T[k] > 2)
   print R[k]
}' file

The explanation:

Code:

{ K[$1,$2,$3]++ }

K is an associative array (associative means keys are strings, not integers) populated with the unique values of the composite key $1 SUBSEP $2 SUBSEP $3, the elements contain the number of occurrences (the count) of these keys.

Code:

$5 ~ /^(14|9[89])$/ && !t[$1,$2,$3,$5]++

For each record that matches:
+ the pattern $5 == 14 or 98 or 99 AND
+ the composite key $1 SUBSEP $2 SUBSEP $3 SUBSEP $5 when seen for the first time (!t[$1,$2,$3,$5]++ is perhaps more readable as t[$1,$2,$3,$5]++ == 0)

Code:

{
T[$1,$2,$3]++
R[$1,$2,$3] = $1 FS $2 FS $3 FS $6
}

Count the occurrences of the key $1 SUBSEP $2 SUBSEP $3 (the T array) and save $1 FS $2 FS $3 FS $6 associated with the key $1 SUBSEP $2 SUBSEP $3 (the R array).

Code:

END {
for (k in K)
 if ((K[k] < 6) && T[k] > 2)
   print R[k]
}

After processing all the input iterate over the K array

Code:

for (k in K)

if the element (the number of occurrences) is less than 6 (a unique section) AND (logical) the element of the T array with the same key
(I used the same keys: $1,$2,$3) is greater than 2 (guarantees that
we have 14, 98 and 99 in the same section), print the element of the R array for that key.


Hope this helps.