script pour établir une bdd Ãà partir d'un tsv avec différent format - retour accueil
git clone git://bebou.netlib.re/laradb
Log | Files | Refs |
date_epoch_debian.sh (1732B)
1 #! /bin/sh
2 # what this script does :
3 # - cleaning the date date so it is under the unix epoch format so that
4 # - it can be used as a value for bc in src/frise.sh
5 # @TODO: apply to both field, and use tsv utils rather than cut to use field header names instead of -f4,5 if
6 # if one day the start_date and end_date aren't anymore the 4th and 5th field of the tsv (however header names must stay persistent
7 # NaN values are defaulted to 1900 01 01 00:00
8
9
10 #start= `mktemp tmp/start.XXXXXXX`
11 #end= `mktemp tmp/end.XXXXXXX `
12 # | xargs date -j +%s # to transform every value into seconds since epoch to bc for the frise
13
14 clean(){
15 cut -f$1 -d" " src/db.tsv |
16 sed '1d;
17 s/-//g; # date 19701210
18 s/^\([0-9]*\)\(.*$\)/\1 #\2/g; # prepare for cleaning with comment
19 s/^\(.{2}\).*/\1/; # finale clean 1970 #comment
20 s/^\([0-9]*\)/\100000000/; #adding the zero we need for the date xargs command to work
21 # second part data cleaning for xargs date
22 s/[0-9]/X/13; #preparation to delete the non-needed zero
23 s/X.[^#]*/ /g; #delete the zero
24 s/^\([12]...\)0000/\10101/; # 197001010000 @TODO for date after 2000**
25 s/^\([12].....\)00/\101/; # 197009010000
26 s/#/ #/; # make the comments a new field #| once there was a pipe
27 s/^0\{8\}/190001010000/; # for NaN gives a date
28 # for debian date add :
29 s/0000 / /; # in format YYYYMMDD '
30 #cut -f1 |
31 #xargs -n1 date -j +%s
32 }
33 clean 4 > tmp/start
34 clean 5 > tmp/end
35 paste -d "\t" tmp/start tmp/end | tee tmp/dateEpoch_final.tsv
36
37 cut -f1 tmp/start |
38 xargs -I XX -d'\n' -n1 date +%s --date="XX" |
39 tee tmp/start_unixepoch
40
41 cut -f1 tmp/end |
42 xargs -I XX -d'\n' -n1 date +%s --date="XX" |
43 tee tmp/end_unixepoch
44
45 paste -d "\t" tmp/start_unixepoch tmp/end_unixepoch | tee tmp/dateEpoch_final.tsv
46
47