nysus has asked for the wisdom of the Perl Monks concerning the following question:

I'm trying to become better at creating nested data structures. I came across this practice problem and decided to try it out. After a couple of hours, I'm stuck trying to figure out how to extract an array of hashes from the data. See the non-functioning second "elsif" conditional in my code below. Here's my attempt (with data):
#!/usr/bin/perl -w use strict; my %alldata = (); while (<DATA>) { chomp; my ($key, $value) = split/\t+/; ## Splits string on one or more t +abs if ( grep { $key eq $_ } qw(ID TITLE GENE CYTOBAND LOCUSLINK CHROMOSOME SCOUNT) ) { ## Procedure for simple hash +es $alldata{$key} = $value; } elsif ($key eq 'EXPRESS') { ## Procedure for simple arrays s/^;+//; ## Gets rid of any leading semicolons $alldata{$key} = [split /;/, $value] } elsif ( grep { $key eq $_ } qw(SEQUENCE PROTSIM) ) { ## Procedur +e for array of hashes my %temphash = (); my @splitvalue = split /;/, $value; foreach my $split2 (@splitvalue) { my ($label, $content) = split/=/, $split2; $temphash{$label} = $content; } $alldata{$key} = push [ %temphash ]; } } __DATA__ ID Hs.22 TITLE transglutaminase 1 (K polypeptide epidermal type I, prote +in-glutamine-gamma-glutamyltransferase) GENE TGM1 CYTOBAND 14q11.2 LOCUSLINK 7051 EXPRESS erwewe;Esophagus;Germ Cell;Larynx;Pancreas;Uterus;colon +;head_neck;uterus CHROMOSOME 14 PROTSIM ORG=Homo sapiens; PROTGI=1070465; PROTID=PIR:TGHUM1; PC +T=100; ALN=816 PROTSIM ORG=Mus musculus; PROTGI=730933; PROTID=SP:Q08189; PCT= +39; ALN=662 PROTSIM ORG=Rattus norvegicus; PROTGI=135697; PROTID=SP:P23606; + PCT=91; ALN=815 SCOUNT 24 SEQUENCE ACC=M62925; NID=g339603; PID=g339604 SEQUENCE ACC=M98447; NID=g186734; PID=g1256959 SEQUENCE ACC=D90287; NID=g219631; PID=g219632 SEQUENCE ACC=M55183; NID=g186789; PID=g186790 SEQUENCE ACC=X57974; NID=g510524; PID=g510525 SEQUENCE ACC=BF155997; NID=g11051180; LID=4808 SEQUENCE ACC=AW083702; NID=g6038854; CLONE=IMAGE:2587766; END=3'; L +ID=728 SEQUENCE ACC=AI652954; NID=g4736933; CLONE=IMAGE:2306445; END=3'; L +ID=698 SEQUENCE ACC=BF155987; NID=g11051170; LID=4808 SEQUENCE ACC=AI239574; NID=g3834971; CLONE=IMAGE:1846343; END=3'; L +ID=600 SEQUENCE ACC=AW265414; NID=g6642230; CLONE=IMAGE:2754263; END=3'; L +ID=1370 SEQUENCE ACC=BE182598; LID=3549 SEQUENCE ACC=AI269864; NID=g3889031; CLONE=IMAGE:2005287; END=3'; L +ID=705 SEQUENCE ACC=AW085789; NID=g6040941; CLONE=IMAGE:2588217; END=3'; L +ID=728 SEQUENCE ACC=BF156003; NID=g11051186; LID=4808 SEQUENCE ACC=AW194040; NID=g6472771; CLONE=IMAGE:2683894; END=3'; L +ID=760 SEQUENCE ACC=AL039214; NID=g5408290; CLONE=DKFZp727C171; END=5'; LI +D=860 SEQUENCE ACC=BE934356; NID=g10460432; LID=4595 SEQUENCE ACC=AW796622; NID=g7848492; LID=2510 SEQUENCE ACC=BE293065; NID=g9175931; CLONE=IMAGE:3349385; END=5'; L +ID=3594 SEQUENCE ACC=AA583940; NID=g2368549; CLONE=IMAGE:1088673; END=3'; L +ID=567 SEQUENCE ACC=NM_000359; NID=g4507474; PID=g4507475 SEQUENCE ACC=BF155992; NID=g11051175; LID=4808 SEQUENCE ACC=BF089798; NID=g10895508; LID=4808

$PM = "Perl Monk's";
$MCF = "Most Clueless Friar";
$nysus = $PM . $MCF;

Replies are listed 'Best First'.
Re: Extracting array of hashes from data
by wog (Curate) on Jun 10, 2001 at 00:32 UTC
    This line seems to be your problem:

    $alldata{$key} = push [ %temphash ];

    you probably really meant to push %temphash on to the array ref in $alldata{$key}. In that case you can reference that array with @{ $alldata{$key} }. Then you can just push a reference to %temphash on top of it:

    push @{ $alldata{$key} }, \%temphash;

    (the array will be automatically created as needed)

    Of course see perldsc and perlref if you haven't already.

    update: Typo fixed. Thanks srawls.

      push @{ $alldata{$key }, \%temphash

      whoops, minor typo, big error; change the above line to this:

      push @{$alldata{$key }}, \%temphash

      or, you can try this:

      $hash{$key} = \%temphash; # $hash{$key} is a reference to %temph +ash $hash{$hey} = \@array; # $hash{$key} is a reference to @array $hash{$key} = {KEY => 'value'}; #$hash{$key} is a reference to an anon +ymous hash $hash{$key} = [1,2,3]; #$hash[$key] is a reference to an anon +ymous array

      Update:Cleaned up code formating a bit

      The 15 year old, freshman programmer,
      Stephen Rawls

      I'm still running into a problem using push @{ $alldata{$key} }, \%temphash;

      For instance, try: print "$alldata{PROTSIM}[2]{PCT}\n"; at the end of the file. You'll get an Use of uninitialized value at line 24, <DATA> chunk 35. error. What's strange is that: print "$alldata{PROTSIM}[2]{ORG}\n"; is AOK.

      Anyone have any ideas?

      I have not read perldsc but I certainly will. Looks useful.

      $PM = "Perl Monk's";
      $MCF = "Most Clueless Friar";
      $nysus = $PM . $MCF;

        Took me a sec, but I got it. In your data, sometimes you have spaces after the ;, and sometimes you don't. After each of your splits (or just the split/;/'s) add  \s* as in:
        my @splitvalue = split /;\s*/, $value;

        The 15 year old, freshman programmer,
        Stephen Rawls

Re: Extracting array of hashes from data
by Abigail (Deacon) on Jun 10, 2001 at 17:42 UTC
    #!/opt/perl/bin/perl -w use strict; my %keys = (map ({$_ => "SCALAR"} qw /ID TITLE GENE CYTOBAND LOCUSLINK CHROMOSOME SCOUNT/), map ({$_ => "ARRAY"} qw /EXPRESS/), map ({$_ => "HASH"} qw /SEQUENCE PROTSIM/)); my %alldata; while (<DATA>) { chomp; my ($key, $value) = split /\t+/ => $_, 2; if ($keys {$key} eq "SCALAR") {$alldata {$key} = $value} elsif ($keys {$key} eq "ARRAY") { $value =~ s/^;+//; $alldata {$key} = [split /;/ => $value]; } elsif ($keys {$key} eq "HASH") { push @{$alldata {$key}} => {map {split /=/} split /;\s*/ => $v +alue}; } else { die "Unknown key $key found in data.\n" } } __DATA__

    -- Abigail

Re: Extracting array of hashes from data
by Zaxo (Archbishop) on Jun 10, 2001 at 01:27 UTC

    You use a good technique for setting $alldata{'EXPRESS'}, splitting on a field delimiter inside array ref brackets:

    $alldata{$key} = [split /;/, $value]

    You can place a SEQUENCE or PROTSYM element in a hash reference the same way:

    # replaces everything in the last elsif(){...} push @{$alldata{$key}}, {split /;|=/, $value};

    This works because a list with an even number of elements can always be treated as a hash. If you want to trim whitespace or something, that can be map'ped inside the braces.

    After Compline,

      Very cool.

      $PM = "Perl Monk's";
      $MCF = "Most Clueless Friar";
      $nysus = $PM . $MCF;

(zdog) Re: Extracting array of hashes from data
by zdog (Priest) on Jun 10, 2001 at 01:33 UTC
    You may be looking to do something like this:

    $alldata{$key} = \%temphash;

    Instead of:

    $alldata{$key} = push [ %temphash ];

    It sets $alldata{$key} as a reference to %temphash. So the content is accessed using $alldata{$key}{$key2}.

    Zenon Zabinski | zdog |