Example:
1905-04-15 name[Sergey Mihoparov] died[2010-07-12] education[school4] jailed[1951-11-02_1959-11-01] moderate-negative single experience[35]
1989-03-29 name[Nikita Zverev] education[school10+university6] marred1- experience[6]
2017-09-14 name[Aleksey Bezborodko] single experience[]
See at moderate-negative charactestics – imagine we don’t care about it. In this case a rule that will consume both these lines may look like
Rule =
Time(string) ' '
^"name[" Name(string) ']' ^' '
??Death(
^"died[" Date(string) ']' ^' '
)
^"education[" Education(string) ']' ^' '
?Jailed(
^"jailed[" Intervals(string) ']' ^' '
)
?( ^"moderate-negative ")
Status(string) ' '
^"experience["
??Experience(
Years(int) ']'
);
Please take a look at optional areas. The area ?Jailed will create a substructure with Valid bool field, which is set to true if it was matched successfully and set it to
false otherwise. Remember though, if you have some numeric taker and conversion failed the whole Rule failed. Use ?? to ignore conversion errors as well.
And so called anonymous area ?( ^"moderate-negative "). It is checking if the rest starts with moderate-negative and passes it in case it is true. Or go to the next action otherwise.
Lines:
Vladimir 37 186 91 15
Mikhail 35 184 76
And we don’t need the last number `Vladimir’ has. So, use this rule
Rule =
Name(string) ' '
Age(int) ' '
Height(int) ' '
Weight(int) ?' '
;
See the ? before the last ' ' – it commands to take everything as Weight right to the end or the first ' '
Let we have something like
2019-09-08T23:47:11.671194+03:00 id=some_id version=1.1.1 coords={"lat":11.11, "lon":22.22} locale=en-EN
and we ' ' won’t work as a bound for coords as we have one in JSON itself. The solution is to use
[] taker, it will store bound itself:
Rule =
Time(string) " id="
ID(string) " version="
Version(string) " coords="
Coords[string] '}' ^" locale="
Locale(string)
;
Let we have two files a.info and b.info:
a.info```a.info a.info 1 2 3 4 12 13 14 15 1 1 1 1
##### `b.info`
1 2 3 4 5 5 4 3 2 1
You see, format in `b.info` clearly extends the one in `a.info` and we would
love to process both in a similar way (with the last column defaulted to some
value in case of `a.info`, most likely to `0`)
How can we achieve this? Tricky, but simple:
### Rules
```perl
A = First(int) ' ' Second(int) ' ' Third(int) ' ' Fourth(int);
B = First(int) ' ' Second(int) ' ' Third(int) ' ' Fourth(int) ` ` Fifth(int);
and compile it
ldetool --package main rule.lde
b.info parses in a simple manner
for bInfo.Next() {
var b B
if ok, err := b.Extract(bInfo.Bytes()); !ok {
if err != nil {
return …
}
return …
}
processInfo(b)
}
parsing a.info:
for aInfo.Next() {
var b B
if ok, err := (*A)(unsafe.Pointer(&b)).Extract(aInfo.Bytes()); !ok {
if err != nil {
return …
}
return …
}
processInfo(b)
}
This will work as memory layout in A looks exactly like the memory layout of the first four fields of B