Fix per-sample statistics

Some of the per-sample stats were assigned to different samples
when `bcftools stats -s/-S`, depending on the order and the sample
count.

Resolves #2469, pull request #2534 with a test added

Author: pd3

NEWS

@@ -33,6 +33,10 @@ Changes affecting specific commands:
       unwanted pair of records, it also led to unexpected behavior when the first
       out-of-window record would change by filtering (#2475)
 
+* bcftools stats
+
+    - Fix per-sample stats when samples provided in different order via -s/-S (#2534)
+
 * bcftools +trio-dnm3
 
     - Add a new TrioDNM model and make it the new default. To prevent confusion, the

test/stats-smpl-order.vcf

@@ -0,0 +1,8 @@
+##fileformat=VCFv4.2
+##contig=<ID=1,length=1000000>
+##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">
+##FORMAT=<ID=DP,Number=1,Type=Integer,Description="Read depth">
+##FORMAT=<ID=AD,Number=R,Type=Integer,Description="Allelic depths">
+#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	A	B
+1	100	.	A	G	60	PASS	.	GT:DP:AD	0/1:10:6,4	0/1:100:1,99
+1	200	.	AT	A	60	PASS	.	GT:DP:AD	0/1:20:10,10	0/1:200:190,10

test/test.pl

@@ -1,6 +1,6 @@
 #!/usr/bin/env perl
 #
-#   Copyright (C) 2012-2025 Genome Research Ltd.
+#   Copyright (C) 2012-2026 Genome Research Ltd.
 #
 #   Author: Petr Danecek <pd3@sanger.ac.uk>
 #
@@ -44,6 +44,8 @@
 run_test(\&test_vcf_idxstats,$opts,in=>'empty',args=>'-n',out=>'empty.idx_count.out');
 run_test(\&test_vcf_check,$opts,in=>'check',out=>'check.chk');
 run_test(\&test_vcf_check_merge,$opts,in=>'check',out=>'check_merge.chk');
+run_test(\&test_vcf_stats,$opts,in=>['stats-smpl-order'],out=>'stats-smpl-order.chk',args=>'-s A,B',pipe=>'sort');
+run_test(\&test_vcf_stats,$opts,in=>['stats-smpl-order'],out=>'stats-smpl-order.chk',args=>'-s B,A',pipe=>'sort');
 run_test(\&test_vcf_stats,$opts,in=>['stats.a','stats.b'],out=>'stats.chk',args=>'-s -');
 run_test(\&test_vcf_stats,$opts,in=>['stats.a','stats.b'],out=>'stats.B.chk',args=>'-s B');
 run_test(\&test_vcf_stats,$opts,in=>['stats.counts'],out=>'stats.counts.chk',args=>'-s -');
@@ -1511,7 +1513,9 @@ sub test_vcf_stats
         bgzip_tabix_vcf($opts,$file);
         $files .= " $$opts{tmp}/$file.vcf.gz";
     }
-    test_cmd($opts,%args,cmd=>"$$opts{bin}/bcftools stats $args{args} $files | grep -v '^#' | grep -v '^ID\t'");
+    my $pipe = qq[grep -v '^#' | grep -v '^ID\t' ];
+    if ( exists($args{pipe}) ) { $pipe .= ' | '. $args{pipe}; }
+    test_cmd($opts,%args,cmd=>"$$opts{bin}/bcftools stats $args{args} $files | $pipe");
 }
 sub test_vcf_merge
 {

vcfstats.c

@@ -1,6 +1,6 @@
 /*  vcfstats.c -- Produces stats which can be plotted using plot-vcfstats.
 
-    Copyright (C) 2012-2025 Genome Research Ltd.
+    Copyright (C) 2012-2026 Genome Research Ltd.
 
     Author: Petr Danecek <pd3@sanger.ac.uk>
 
@@ -1103,8 +1103,10 @@ static void do_sample_stats(args_t *args, stats_t *stats, bcf_sr_t *reader, int
     int is;
     for (is=0; is<args->files->n_smpl; is++)
     {
+        int ismpl = reader->samples[is];    // VCF column index for this sample after possible reordering
+
         // Determine depth
-        int dp = calc_sample_depth(args,is,ad_fmt_ptr,dp_fmt_ptr);
+        int dp = calc_sample_depth(args,ismpl,ad_fmt_ptr,dp_fmt_ptr);
         if ( dp>0 )
         {
             (*idist(&stats->dp, dp))++;
@@ -1116,7 +1118,7 @@ static void do_sample_stats(args_t *args, stats_t *stats, bcf_sr_t *reader, int
         int ial, jal, gt=GT_UNKN;
         if ( gt_fmt_ptr )
         {
-            gt = bcf_gt_type(gt_fmt_ptr, reader->samples[is], &ial, &jal);
+            gt = bcf_gt_type(gt_fmt_ptr, ismpl, &ial, &jal);
             sample_gt_stats(args,stats,line,is,gt,ial,jal);
         }
 
@@ -1126,12 +1128,12 @@ static void do_sample_stats(args_t *args, stats_t *stats, bcf_sr_t *reader, int
             float iad = 0, jad = 0;
             if ( gt==GT_UNKN )    // GT not available
             {
-                iad = get_ad(line,ad_fmt_ptr,is,&ial);
+                iad = get_ad(line,ad_fmt_ptr,ismpl,&ial);
             }
             else if ( gt!=GT_UNKN )
             {
-                iad = ial==0 ? 0 : get_iad(line,ad_fmt_ptr,is,ial);
-                jad = jal==0 ? 0 : get_iad(line,ad_fmt_ptr,is,jal);
+                iad = ial==0 ? 0 : get_iad(line,ad_fmt_ptr,ismpl,ial);
+                jad = jal==0 ? 0 : get_iad(line,ad_fmt_ptr,ismpl,jal);
             }
             if ( iad )
             {