#include <stdlib.h>
#include <ctype.h>
#include <assert.h>
+#include <errno.h>
#include <stdio.h>
#include <string.h>
#include <unistd.h>
KSORT_INIT(heap, heap1_t, heap_lt)
+static void swap_header_targets(bam_header_t *h1, bam_header_t *h2)
+{
+ bam_header_t t;
+ t.n_targets = h1->n_targets, h1->n_targets = h2->n_targets, h2->n_targets = t.n_targets;
+ t.target_name = h1->target_name, h1->target_name = h2->target_name, h2->target_name = t.target_name;
+ t.target_len = h1->target_len, h1->target_len = h2->target_len, h2->target_len = t.target_len;
+}
+
static void swap_header_text(bam_header_t *h1, bam_header_t *h2)
{
int tempi;
if (headers) {
tamFile fpheaders = sam_open(headers);
if (fpheaders == 0) {
- fprintf(stderr, "[bam_merge_core] Cannot open file `%s'. Continue anyway.\n", headers);
- } else {
- hheaders = sam_header_read(fpheaders);
- sam_close(fpheaders);
+ const char *message = strerror(errno);
+ fprintf(stderr, "[bam_merge_core] cannot open '%s': %s\n", headers, message);
+ return -1;
}
+ hheaders = sam_header_read(fpheaders);
+ sam_close(fpheaders);
}
g_is_by_qname = by_qname;
return -1;
}
hin = bam_header_read(fp[i]);
- if (i == 0) { // the first SAM
+ if (i == 0) { // the first BAM
hout = hin;
- if (hheaders) {
- // If the text headers to be swapped in include any @SQ headers,
- // check that they are consistent with the existing binary list
- // of reference information.
- if (hheaders->n_targets > 0) {
- if (hout->n_targets != hheaders->n_targets)
- fprintf(stderr, "[bam_merge_core] number of @SQ headers in `%s' differs from number of target sequences", headers);
- for (j = 0; j < hout->n_targets; ++j)
- if (strcmp(hout->target_name[j], hheaders->target_name[j]) != 0)
- fprintf(stderr, "[bam_merge_core] @SQ header '%s' in '%s' differs from target sequence", hheaders->target_name[j], headers);
- }
- swap_header_text(hout, hheaders);
- bam_header_destroy(hheaders);
- hheaders = NULL;
- }
} else { // validate multiple baf
- if (hout->n_targets != hin->n_targets) {
- fprintf(stderr, "[bam_merge_core] file '%s' has different number of target sequences. Abort!\n", fn[i]);
- exit(1);
- }
- for (j = 0; j < hout->n_targets; ++j) {
- if (strcmp(hout->target_name[j], hin->target_name[j])) {
- fprintf(stderr, "[bam_merge_core] different target sequence name: '%s' != '%s' in file '%s'. Abort!\n",
+ int min_n_targets = hout->n_targets;
+ if (hin->n_targets < min_n_targets) min_n_targets = hin->n_targets;
+
+ for (j = 0; j < min_n_targets; ++j)
+ if (strcmp(hout->target_name[j], hin->target_name[j]) != 0) {
+ fprintf(stderr, "[bam_merge_core] different target sequence name: '%s' != '%s' in file '%s'\n",
hout->target_name[j], hin->target_name[j], fn[i]);
- exit(1);
+ return -1;
}
+
+ // If this input file has additional target reference sequences,
+ // add them to the headers to be output
+ if (hin->n_targets > hout->n_targets) {
+ swap_header_targets(hout, hin);
+ // FIXME Possibly we should also create @SQ text headers
+ // for the newly added reference sequences
}
+
bam_header_destroy(hin);
}
}
+ if (hheaders) {
+ // If the text headers to be swapped in include any @SQ headers,
+ // check that they are consistent with the existing binary list
+ // of reference information.
+ if (hheaders->n_targets > 0) {
+ if (hout->n_targets != hheaders->n_targets) {
+ fprintf(stderr, "[bam_merge_core] number of @SQ headers in '%s' differs from number of target sequences\n", headers);
+ if (!reg) return -1;
+ }
+ for (j = 0; j < hout->n_targets; ++j)
+ if (strcmp(hout->target_name[j], hheaders->target_name[j]) != 0) {
+ fprintf(stderr, "[bam_merge_core] @SQ header '%s' in '%s' differs from target sequence\n", hheaders->target_name[j], headers);
+ if (!reg) return -1;
+ }
+ }
+
+ swap_header_text(hout, hheaders);
+ bam_header_destroy(hheaders);
+ }
+
if (reg) {
int tid, beg, end;
if (bam_parse_region(hout, reg, &tid, &beg, &end) < 0) {
ks_heapmake(heap, n, heap);
while (heap->pos != HEAP_EMPTY) {
bam1_t *b = heap->b;
- if ((flag & MERGE_RG) && bam_aux_get(b, "RG") == 0)
+ if (flag & MERGE_RG) {
+ uint8_t *rg = bam_aux_get(b, "RG");
+ if (rg) bam_aux_del(b, rg);
bam_aux_append(b, "RG", 'Z', RG_len[heap->i] + 1, (uint8_t*)RG[heap->i]);
+ }
bam_write1_core(fpout, &b->core, b->data_len, b->data);
if ((j = bam_iter_read(fp[heap->i], iter[heap->i], b)) >= 0) {
heap->pos = ((uint64_t)b->core.tid<<32) | (uint32_t)b->core.pos<<1 | bam1_strand(b);
int bam_merge(int argc, char *argv[])
{
- int c, is_by_qname = 0, flag = 0;
+ int c, is_by_qname = 0, flag = 0, ret = 0;
char *fn_headers = NULL, *reg = 0;
while ((c = getopt(argc, argv, "h:nruR:")) >= 0) {
fprintf(stderr, " the header dictionary in merging.\n\n");
return 1;
}
- bam_merge_core(is_by_qname, argv[optind], fn_headers, argc - optind - 1, argv + optind + 1, flag, reg);
+ if (bam_merge_core(is_by_qname, argv[optind], fn_headers, argc - optind - 1, argv + optind + 1, flag, reg) < 0) ret = 1;
free(reg);
free(fn_headers);
- return 0;
+ return ret;
}
typedef bam1_t *bam1_p;
projects / samtools.git / blobdiff