static void main_usage_separator_options(FILE* o, char* argv0) {
fprintf(o, " --rs --irs --ors Record separators, e.g. 'lf' or '\\r\\n'\n");
fprintf(o, " --fs --ifs --ofs --repifs Field separators, e.g. comma\n");
fprintf(o, " --ps --ips --ops Pair separators, e.g. equals sign\n");
fprintf(o, " Notes:\n");
fprintf(o, " * IPS/OPS are only used for DKVP and XTAB formats, since only in these formats\n");
fprintf(o, " do key-value pairs appear juxtaposed.\n");
fprintf(o, " * IRS/ORS are ignored for XTAB format. Nominally IFS and OFS are newlines;\n");
fprintf(o, " XTAB records are separated by two or more consecutive IFS/OFS -- i.e.\n");
fprintf(o, " a blank line.\n");
fprintf(o, " * OFS must be single-character for PPRINT format. This is because it is used\n");
fprintf(o, " with repetition for alignment; multi-character separators would make\n");
fprintf(o, " alignment impossible.\n");
fprintf(o, " * OPS may be multi-character for XTAB format, in which case alignment is\n");
fprintf(o, " disabled.\n");
fprintf(o, " * DKVP, NIDX, CSVLITE, PPRINT, and XTAB formats are intended to handle\n");
fprintf(o, " platform-native text data. In particular, this means LF line-terminators\n");
fprintf(o, " by default on Linux/OSX. You can use \"--dkvp --rs crlf\" for\n");
fprintf(o, " CRLF-terminated DKVP files, and so on.\n");
fprintf(o, " * CSV is intended to handle RFC-4180-compliant data. In particular, this means\n");
fprintf(o, " it uses CRLF line-terminators by default. You can use \"--csv --rs lf\" for\n");
fprintf(o, " Linux-native CSV files.\n");
fprintf(o, " * All RS/FS/PS options are ignored for JSON format: JSON doesn't allow\n");
fprintf(o, " changing these.\n");
fprintf(o, " * You can specify separators in any of the following ways, shown by example:\n");
fprintf(o, " - Type them out, quoting as necessary for shell escapes, e.g.\n");
fprintf(o, " \"--fs '|' --ips :\"\n");
fprintf(o, " - C-style escape sequences, e.g. \"--rs '\\r\\n' --fs '\\t'\".\n");
fprintf(o, " - To avoid backslashing, you can use any of the following names:\n");
fprintf(o, " ");
lhmss_t* pmap = get_desc_to_chars_map();
for (lhmsse_t* pe = pmap->phead; pe != NULL; pe = pe->pnext) {
fprintf(o, " %s", pe->key);
}
fprintf(o, "\n");
fprintf(o, " * Default separators by format:\n");
fprintf(o, " %-12s %-8s %-8s %s\n", "File format", "RS", "FS", "PS");
lhmss_t* default_rses = get_default_rses();
lhmss_t* default_fses = get_default_fses();
lhmss_t* default_pses = get_default_pses();
for (lhmsse_t* pe = default_rses->phead; pe != NULL; pe = pe->pnext) {
char* filefmt = pe->key;
char* rs = pe->value;
char* fs = lhmss_get(default_fses, filefmt);
char* ps = lhmss_get(default_pses, filefmt);
fprintf(o, " %-12s %-8s %-8s %s\n", filefmt, rebackslash(rs), rebackslash(fs), rebackslash(ps));
}
}
static char* sep_from_arg(char* arg, char* argv0) {
char* chars = lhmss_get(get_desc_to_chars_map(), arg);
if (chars != NULL) // E.g. crlf
return chars;
else // E.g. '\r\n'
return mlr_unbackslash(arg);
}
static void stats1_percentile_emit(void* pvstate, char* value_field_name, char* stats1_acc_name, int copy_data, lrec_t* poutrec) {
stats1_percentile_state_t* pstate = pvstate;
double p;
(void)sscanf(stats1_acc_name, "p%lf", &p); // Assuming this was range-checked earlier on to be in [0,100].
mv_t v = percentile_keeper_emit(pstate->ppercentile_keeper, p);
char* s = mv_alloc_format_val(&v);
// For this type, one accumulator tracks many stats1_names, but a single value_field_name.
char* output_field_name = lhmss_get(pstate->poutput_field_names, stats1_acc_name);
if (output_field_name == NULL) {
output_field_name = mlr_paste_3_strings(value_field_name, "_", stats1_acc_name);
lhmss_put(pstate->poutput_field_names, mlr_strdup_or_die(stats1_acc_name),
output_field_name, FREE_ENTRY_KEY|FREE_ENTRY_VALUE);
}
lrec_put(poutrec, mlr_strdup_or_die(output_field_name), s, FREE_ENTRY_KEY|FREE_ENTRY_VALUE);
}
// ----------------------------------------------------------------
cli_opts_t* parse_command_line(int argc, char** argv) {
cli_opts_t* popts = mlr_malloc_or_die(sizeof(cli_opts_t));
memset(popts, 0, sizeof(*popts));
popts->irs = NULL;
popts->ifs = NULL;
popts->ips = NULL;
popts->allow_repeat_ifs = NEITHER_TRUE_NOR_FALSE;
popts->allow_repeat_ips = NEITHER_TRUE_NOR_FALSE;
popts->use_implicit_csv_header = FALSE;
popts->headerless_csv_output = FALSE;
popts->ors = NULL;
popts->ofs = NULL;
popts->ops = NULL;
popts->right_justify_xtab_value = FALSE;
popts->stack_json_output_vertically = FALSE;
popts->wrap_json_output_in_outer_list = FALSE;
popts->quote_json_values_always = FALSE;
popts->json_flatten_separator = DEFAULT_JSON_FLATTEN_SEPARATOR;
popts->ofmt = DEFAULT_OFMT;
popts->oquoting = DEFAULT_OQUOTING;
popts->plrec_reader = NULL;
popts->plrec_writer = NULL;
popts->prepipe = NULL;
popts->filenames = NULL;
popts->ifile_fmt = "dkvp";
popts->ofile_fmt = "dkvp";
popts->use_mmap_for_read = TRUE;
int left_align_pprint = TRUE;
int have_rand_seed = FALSE;
unsigned rand_seed = 0;
int argi = 1;
for (; argi < argc; argi++) {
if (argv[argi][0] != '-') {
break;
} else if (streq(argv[argi], "--version")) {
#ifdef HAVE_CONFIG_H
printf("Miller %s\n", PACKAGE_VERSION);
#else
printf("Miller %s\n", MLR_VERSION);
#endif // HAVE_CONFIG_H
exit(0);
} else if (streq(argv[argi], "-h")) {
main_usage(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--help")) {
main_usage(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--help-all-verbs")) {
usage_all_verbs(argv[0]);
} else if (streq(argv[argi], "--list-all-verbs") || streq(argv[argi], "-l")) {
list_all_verbs(stdout, "");
exit(0);
} else if (streq(argv[argi], "--list-all-verbs-raw")) {
list_all_verbs_raw(stdout);
exit(0);
} else if (streq(argv[argi], "--list-all-functions-raw")) {
lrec_evaluator_list_all_functions_raw(stdout);
exit(0);
} else if (streq(argv[argi], "--help-all-functions") || streq(argv[argi], "-f")) {
lrec_evaluator_function_usage(stdout, NULL);
exit(0);
} else if (streq(argv[argi], "--help-function") || streq(argv[argi], "--hf")) {
check_arg_count(argv, argi, argc, 2);
lrec_evaluator_function_usage(stdout, argv[argi+1]);
exit(0);
// main-usage subsections, individually accessible for the benefit of
// the manpage-autogenerator
} else if (streq(argv[argi], "--usage-synopsis")) {
main_usage_synopsis(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--usage-examples")) {
main_usage_examples(stdout, argv[0], "");
exit(0);
} else if (streq(argv[argi], "--usage-list-all-verbs")) {
list_all_verbs(stdout, "");
exit(0);
} else if (streq(argv[argi], "--usage-help-options")) {
main_usage_help_options(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--usage-functions")) {
main_usage_functions(stdout, argv[0], "");
exit(0);
} else if (streq(argv[argi], "--usage-data-format-examples")) {
main_usage_data_format_examples(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--usage-data-format-options")) {
main_usage_data_format_options(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--usage-compressed-data-options")) {
main_usage_compressed_data_options(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--usage-separator-options")) {
main_usage_separator_options(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--usage-csv-options")) {
main_usage_csv_options(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--usage-double-quoting")) {
main_usage_double_quoting(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--usage-numerical-formatting")) {
main_usage_numerical_formatting(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--usage-other-options")) {
main_usage_other_options(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--usage-then-chaining")) {
main_usage_then_chaining(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--usage-see-also")) {
main_usage_see_also(stdout, argv[0]);
exit(0);
} else if (streq(argv[argi], "--rs")) {
check_arg_count(argv, argi, argc, 2);
popts->ors = sep_from_arg(argv[argi+1], argv[0]);
popts->irs = sep_from_arg(argv[argi+1], argv[0]);
argi++;
} else if (streq(argv[argi], "--irs")) {
check_arg_count(argv, argi, argc, 2);
popts->irs = sep_from_arg(argv[argi+1], argv[0]);
argi++;
} else if (streq(argv[argi], "--ors")) {
check_arg_count(argv, argi, argc, 2);
popts->ors = sep_from_arg(argv[argi+1], argv[0]);
argi++;
} else if (streq(argv[argi], "--fs")) {
check_arg_count(argv, argi, argc, 2);
popts->ofs = sep_from_arg(argv[argi+1], argv[0]);
popts->ifs = sep_from_arg(argv[argi+1], argv[0]);
argi++;
} else if (streq(argv[argi], "--ifs")) {
check_arg_count(argv, argi, argc, 2);
popts->ifs = sep_from_arg(argv[argi+1], argv[0]);
argi++;
} else if (streq(argv[argi], "--ofs")) {
check_arg_count(argv, argi, argc, 2);
popts->ofs = sep_from_arg(argv[argi+1], argv[0]);
argi++;
} else if (streq(argv[argi], "--repifs")) {
popts->allow_repeat_ifs = TRUE;
} else if (streq(argv[argi], "--implicit-csv-header")) {
popts->use_implicit_csv_header = TRUE;
} else if (streq(argv[argi], "--headerless-csv-output")) {
popts->headerless_csv_output = TRUE;
} else if (streq(argv[argi], "-p")) {
popts->ifile_fmt = "nidx";
popts->ofile_fmt = "nidx";
popts->ifs = " ";
popts->ofs = " ";
popts->allow_repeat_ifs = TRUE;
} else if (streq(argv[argi], "--ps")) {
check_arg_count(argv, argi, argc, 2);
popts->ops = sep_from_arg(argv[argi+1], argv[0]);
popts->ips = sep_from_arg(argv[argi+1], argv[0]);
argi++;
} else if (streq(argv[argi], "--ips")) {
check_arg_count(argv, argi, argc, 2);
popts->ips = sep_from_arg(argv[argi+1], argv[0]);
argi++;
} else if (streq(argv[argi], "--ops")) {
check_arg_count(argv, argi, argc, 2);
popts->ops = sep_from_arg(argv[argi+1], argv[0]);
argi++;
} else if (streq(argv[argi], "--xvright")) {
popts->right_justify_xtab_value = TRUE;
} else if (streq(argv[argi], "--jvstack")) {
popts->stack_json_output_vertically = TRUE;
} else if (streq(argv[argi], "--jlistwrap")) {
popts->wrap_json_output_in_outer_list = TRUE;
} else if (streq(argv[argi], "--jquoteall")) {
popts->quote_json_values_always = TRUE;
} else if (streq(argv[argi], "--jflatsep")) {
check_arg_count(argv, argi, argc, 2);
popts->json_flatten_separator = sep_from_arg(argv[argi+1], argv[0]);
argi++;
} else if (streq(argv[argi], "--csv")) { popts->ifile_fmt = popts->ofile_fmt = "csv";
} else if (streq(argv[argi], "--icsv")) { popts->ifile_fmt = "csv";
} else if (streq(argv[argi], "--ocsv")) { popts->ofile_fmt = "csv";
} else if (streq(argv[argi], "--csvlite")) { popts->ifile_fmt = popts->ofile_fmt = "csvlite";
} else if (streq(argv[argi], "--icsvlite")) { popts->ifile_fmt = "csvlite";
} else if (streq(argv[argi], "--ocsvlite")) { popts->ofile_fmt = "csvlite";
} else if (streq(argv[argi], "--dkvp")) { popts->ifile_fmt = popts->ofile_fmt = "dkvp";
} else if (streq(argv[argi], "--idkvp")) { popts->ifile_fmt = "dkvp";
} else if (streq(argv[argi], "--odkvp")) { popts->ofile_fmt = "dkvp";
} else if (streq(argv[argi], "--json")) { popts->ifile_fmt = popts->ofile_fmt = "json";
} else if (streq(argv[argi], "--ijson")) { popts->ifile_fmt = "json";
} else if (streq(argv[argi], "--ojson")) { popts->ofile_fmt = "json";
} else if (streq(argv[argi], "--nidx")) { popts->ifile_fmt = popts->ofile_fmt = "nidx";
} else if (streq(argv[argi], "--inidx")) { popts->ifile_fmt = "nidx";
} else if (streq(argv[argi], "--onidx")) { popts->ofile_fmt = "nidx";
} else if (streq(argv[argi], "--xtab")) { popts->ifile_fmt = popts->ofile_fmt = "xtab";
} else if (streq(argv[argi], "--ixtab")) { popts->ifile_fmt = "xtab";
} else if (streq(argv[argi], "--oxtab")) { popts->ofile_fmt = "xtab";
} else if (streq(argv[argi], "--ipprint")) {
popts->ifile_fmt = "csvlite";
popts->ifs = " ";
popts->allow_repeat_ifs = TRUE;
} else if (streq(argv[argi], "--opprint")) {
popts->ofile_fmt = "pprint";
} else if (streq(argv[argi], "--pprint")) {
popts->ifile_fmt = "csvlite";
popts->ifs = " ";
popts->allow_repeat_ifs = TRUE;
popts->ofile_fmt = "pprint";
} else if (streq(argv[argi], "--right")) {
left_align_pprint = FALSE;
} else if (streq(argv[argi], "--ofmt")) {
check_arg_count(argv, argi, argc, 2);
popts->ofmt = argv[argi+1];
argi++;
} else if (streq(argv[argi], "--quote-all")) { popts->oquoting = QUOTE_ALL;
} else if (streq(argv[argi], "--quote-none")) { popts->oquoting = QUOTE_NONE;
} else if (streq(argv[argi], "--quote-minimal")) { popts->oquoting = QUOTE_MINIMAL;
} else if (streq(argv[argi], "--quote-numeric")) { popts->oquoting = QUOTE_NUMERIC;
} else if (streq(argv[argi], "--mmap")) {
popts->use_mmap_for_read = TRUE;
} else if (streq(argv[argi], "--no-mmap")) {
popts->use_mmap_for_read = FALSE;
} else if (streq(argv[argi], "--seed")) {
check_arg_count(argv, argi, argc, 2);
if (sscanf(argv[argi+1], "0x%x", &rand_seed) == 1) {
have_rand_seed = TRUE;
} else if (sscanf(argv[argi+1], "%u", &rand_seed) == 1) {
have_rand_seed = TRUE;
} else {
main_usage(stderr, argv[0]);
exit(1);
}
argi++;
} else if (streq(argv[argi], "--prepipe")) {
check_arg_count(argv, argi, argc, 2);
popts->prepipe = argv[argi+1];
popts->use_mmap_for_read = FALSE;
argi++;
} else {
usage_unrecognized_verb(argv[0], argv[argi]);
}
}
lhmss_t* default_rses = get_default_rses();
lhmss_t* default_fses = get_default_fses();
lhmss_t* default_pses = get_default_pses();
lhmsi_t* default_repeat_ifses = get_default_repeat_ifses();
lhmsi_t* default_repeat_ipses = get_default_repeat_ipses();
if (popts->irs == NULL)
popts->irs = lhmss_get(default_rses, popts->ifile_fmt);
if (popts->ifs == NULL)
popts->ifs = lhmss_get(default_fses, popts->ifile_fmt);
if (popts->ips == NULL)
popts->ips = lhmss_get(default_pses, popts->ifile_fmt);
if (popts->allow_repeat_ifs == NEITHER_TRUE_NOR_FALSE)
popts->allow_repeat_ifs = lhmsi_get(default_repeat_ifses, popts->ifile_fmt);
if (popts->allow_repeat_ips == NEITHER_TRUE_NOR_FALSE)
popts->allow_repeat_ips = lhmsi_get(default_repeat_ipses, popts->ifile_fmt);
if (popts->ors == NULL)
popts->ors = lhmss_get(default_rses, popts->ofile_fmt);
if (popts->ofs == NULL)
popts->ofs = lhmss_get(default_fses, popts->ofile_fmt);
if (popts->ops == NULL)
popts->ops = lhmss_get(default_pses, popts->ofile_fmt);
if (popts->irs == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->ifs == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->ips == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->allow_repeat_ifs == NEITHER_TRUE_NOR_FALSE) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->allow_repeat_ips == NEITHER_TRUE_NOR_FALSE) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->ors == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->ofs == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->ops == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (streq(popts->ofile_fmt, "pprint") && strlen(popts->ofs) != 1) {
fprintf(stderr, "%s: OFS for PPRINT format must be single-character; got \"%s\".\n",
argv[0], popts->ofs);
return NULL;
}
if (streq(popts->ofile_fmt, "dkvp"))
popts->plrec_writer = lrec_writer_dkvp_alloc(popts->ors, popts->ofs, popts->ops);
else if (streq(popts->ofile_fmt, "json"))
popts->plrec_writer = lrec_writer_json_alloc(popts->stack_json_output_vertically,
popts->wrap_json_output_in_outer_list, popts->quote_json_values_always, popts->json_flatten_separator);
else if (streq(popts->ofile_fmt, "csv"))
popts->plrec_writer = lrec_writer_csv_alloc(popts->ors, popts->ofs, popts->oquoting,
popts->headerless_csv_output);
else if (streq(popts->ofile_fmt, "csvlite"))
popts->plrec_writer = lrec_writer_csvlite_alloc(popts->ors, popts->ofs, popts->headerless_csv_output);
else if (streq(popts->ofile_fmt, "nidx"))
popts->plrec_writer = lrec_writer_nidx_alloc(popts->ors, popts->ofs);
else if (streq(popts->ofile_fmt, "xtab"))
popts->plrec_writer = lrec_writer_xtab_alloc(popts->ofs, popts->ops, popts->right_justify_xtab_value);
else if (streq(popts->ofile_fmt, "pprint"))
popts->plrec_writer = lrec_writer_pprint_alloc(popts->ors, popts->ofs[0], left_align_pprint);
else {
main_usage(stderr, argv[0]);
exit(1);
}
if ((argc - argi) < 1) {
main_usage(stderr, argv[0]);
exit(1);
}
popts->pmapper_list = sllv_alloc();
while (TRUE) {
check_arg_count(argv, argi, argc, 1);
char* verb = argv[argi];
mapper_setup_t* pmapper_setup = look_up_mapper_setup(verb);
if (pmapper_setup == NULL) {
fprintf(stderr, "%s: verb \"%s\" not found. Please use \"%s --help\" for a list.\n",
argv[0], verb, argv[0]);
exit(1);
}
if ((argc - argi) >= 2) {
if (streq(argv[argi+1], "-h") || streq(argv[argi+1], "--help")) {
pmapper_setup->pusage_func(stdout, argv[0], verb);
exit(0);
}
}
// It's up to the parse func to print its usage on CLI-parse failure.
mapper_t* pmapper = pmapper_setup->pparse_func(&argi, argc, argv);
if (pmapper == NULL) {
exit(1);
}
sllv_append(popts->pmapper_list, pmapper);
if (argi >= argc || !streq(argv[argi], "then"))
break;
argi++;
}
popts->filenames = &argv[argi];
// No filenames means read from standard input, and standard input cannot be mmapped.
if (argi == argc)
popts->use_mmap_for_read = FALSE;
popts->plrec_reader = lrec_reader_alloc(popts->ifile_fmt, popts->use_mmap_for_read,
popts->irs, popts->ifs, popts->allow_repeat_ifs, popts->ips, popts->allow_repeat_ips,
popts->use_implicit_csv_header, popts->json_flatten_separator);
if (popts->plrec_reader == NULL) {
main_usage(stderr, argv[0]);
exit(1);
}
if (have_rand_seed) {
mtrand_init(rand_seed);
} else {
mtrand_init_default();
}
return popts;
}
// ----------------------------------------------------------------
static char* test_lhmss() {
lhmss_t *pmap = lhmss_alloc();
mu_assert_lf(pmap->num_occupied == 0);
mu_assert_lf(!lhmss_has_key(pmap, "w")); mu_assert_lf(lhmss_get(pmap, "w") == NULL);
mu_assert_lf(!lhmss_has_key(pmap, "x")); mu_assert_lf(lhmss_get(pmap, "x") == NULL);
mu_assert_lf(!lhmss_has_key(pmap, "y")); mu_assert_lf(lhmss_get(pmap, "y") == NULL);
mu_assert_lf(!lhmss_has_key(pmap, "z")); mu_assert_lf(lhmss_get(pmap, "z") == NULL);
mu_assert_lf(lhmss_check_counts(pmap));
lhmss_put(pmap, "x", "3");
mu_assert_lf(pmap->num_occupied == 1);
mu_assert_lf(!lhmss_has_key(pmap, "w")); mu_assert_lf(lhmss_get(pmap, "w") == NULL);
mu_assert_lf( lhmss_has_key(pmap, "x")); mu_assert_lf(streq(lhmss_get(pmap, "x"), "3"));
mu_assert_lf(!lhmss_has_key(pmap, "y")); mu_assert_lf(lhmss_get(pmap, "y") == NULL);
mu_assert_lf(!lhmss_has_key(pmap, "z")); mu_assert_lf(lhmss_get(pmap, "z") == NULL);
mu_assert_lf(lhmss_check_counts(pmap));
lhmss_put(pmap, "y", "5");
mu_assert_lf(pmap->num_occupied == 2);
mu_assert_lf(!lhmss_has_key(pmap, "w")); mu_assert_lf(lhmss_get(pmap, "w") == NULL);
mu_assert_lf( lhmss_has_key(pmap, "x")); mu_assert_lf(streq(lhmss_get(pmap, "x"), "3"));
mu_assert_lf( lhmss_has_key(pmap, "y")); mu_assert_lf(streq(lhmss_get(pmap, "y"), "5"));
mu_assert_lf(!lhmss_has_key(pmap, "z")); mu_assert_lf(lhmss_get(pmap, "z") == NULL);
mu_assert_lf(lhmss_check_counts(pmap));
lhmss_put(pmap, "x", "4");
mu_assert_lf(pmap->num_occupied == 2);
mu_assert_lf(!lhmss_has_key(pmap, "w")); mu_assert_lf(lhmss_get(pmap, "w") == NULL);
mu_assert_lf( lhmss_has_key(pmap, "x")); mu_assert_lf(streq(lhmss_get(pmap, "x"), "4"));
mu_assert_lf( lhmss_has_key(pmap, "y")); mu_assert_lf(streq(lhmss_get(pmap, "y"), "5"));
mu_assert_lf(!lhmss_has_key(pmap, "z")); mu_assert_lf(lhmss_get(pmap, "z") == NULL);
mu_assert_lf(lhmss_check_counts(pmap));
lhmss_put(pmap, "z", "7");
mu_assert_lf(pmap->num_occupied == 3);
mu_assert_lf(!lhmss_has_key(pmap, "w")); mu_assert_lf(lhmss_get(pmap, "w") == NULL);
mu_assert_lf( lhmss_has_key(pmap, "x")); mu_assert_lf(streq(lhmss_get(pmap, "x"), "4"));
mu_assert_lf( lhmss_has_key(pmap, "y")); mu_assert_lf(streq(lhmss_get(pmap, "y"), "5"));
mu_assert_lf( lhmss_has_key(pmap, "z")); mu_assert_lf(streq(lhmss_get(pmap, "z"), "7"));
mu_assert_lf(lhmss_check_counts(pmap));
lhmss_remove(pmap, "y");
mu_assert_lf(pmap->num_occupied == 2);
mu_assert_lf(!lhmss_has_key(pmap, "w")); mu_assert_lf(lhmss_get(pmap, "w") == NULL);
mu_assert_lf( lhmss_has_key(pmap, "x")); mu_assert_lf(streq(lhmss_get(pmap, "x"), "4"));
mu_assert_lf(!lhmss_has_key(pmap, "y")); mu_assert_lf(lhmss_get(pmap, "y") == NULL);
mu_assert_lf( lhmss_has_key(pmap, "z")); mu_assert_lf(streq(lhmss_get(pmap, "z"), "7"));
mu_assert_lf(lhmss_check_counts(pmap));
lhmss_free(pmap);
return NULL;
}
// ----------------------------------------------------------------
cli_opts_t* parse_command_line(int argc, char** argv) {
cli_opts_t* popts = mlr_malloc_or_die(sizeof(cli_opts_t));
memset(popts, 0, sizeof(*popts));
popts->irs = NULL;
popts->ifs = NULL;
popts->ips = NULL;
popts->allow_repeat_ifs = NEITHER_TRUE_NOR_FALSE;
popts->allow_repeat_ips = NEITHER_TRUE_NOR_FALSE;
popts->ors = NULL;
popts->ofs = NULL;
popts->ops = NULL;
popts->ofmt = DEFAULT_OFMT;
popts->oquoting = DEFAULT_OQUOTING;
popts->plrec_reader = NULL;
popts->plrec_writer = NULL;
popts->filenames = NULL;
popts->ifile_fmt = "dkvp";
popts->ofile_fmt = "dkvp";
popts->use_mmap_for_read = TRUE;
int left_align_pprint = TRUE;
int have_rand_seed = FALSE;
unsigned rand_seed = 0;
int argi = 1;
for (; argi < argc; argi++) {
if (argv[argi][0] != '-')
break;
else if (streq(argv[argi], "--version")) {
#ifdef HAVE_CONFIG_H
printf("Miller version >= %s.\n", PACKAGE_VERSION);
#else
printf("Miller version >= %s.\n", MLR_VERSION);
#endif // HAVE_CONFIG_H
exit(0);
} else if (streq(argv[argi], "-h"))
main_usage(argv[0], 0);
else if (streq(argv[argi], "--help"))
main_usage(argv[0], 0);
else if (streq(argv[argi], "--help-all-verbs"))
usage_all_verbs(argv[0]);
else if (streq(argv[argi], "--help-all-functions") || streq(argv[argi], "-f")) {
lrec_evaluator_function_usage(stdout, NULL);
exit(0);
}
else if (streq(argv[argi], "--help-function") || streq(argv[argi], "--hf")) {
check_arg_count(argv, argi, argc, 2);
lrec_evaluator_function_usage(stdout, argv[argi+1]);
exit(0);
}
else if (streq(argv[argi], "--rs")) {
check_arg_count(argv, argi, argc, 2);
popts->ors = sep_from_arg(argv[argi+1], argv[0]);
popts->irs = sep_from_arg(argv[argi+1], argv[0]);
argi++;
}
else if (streq(argv[argi], "--irs")) {
check_arg_count(argv, argi, argc, 2);
popts->irs = sep_from_arg(argv[argi+1], argv[0]);
argi++;
}
else if (streq(argv[argi], "--ors")) {
check_arg_count(argv, argi, argc, 2);
popts->ors = sep_from_arg(argv[argi+1], argv[0]);
argi++;
}
else if (streq(argv[argi], "--fs")) {
check_arg_count(argv, argi, argc, 2);
popts->ofs = sep_from_arg(argv[argi+1], argv[0]);
popts->ifs = sep_from_arg(argv[argi+1], argv[0]);
argi++;
}
else if (streq(argv[argi], "--ifs")) {
check_arg_count(argv, argi, argc, 2);
popts->ifs = sep_from_arg(argv[argi+1], argv[0]);
argi++;
}
else if (streq(argv[argi], "--ofs")) {
check_arg_count(argv, argi, argc, 2);
popts->ofs = sep_from_arg(argv[argi+1], argv[0]);
argi++;
}
else if (streq(argv[argi], "--repifs")) {
popts->allow_repeat_ifs = TRUE;
}
else if (streq(argv[argi], "-p")) {
popts->ifile_fmt = "nidx";
popts->ofile_fmt = "nidx";
popts->ifs = " ";
popts->ofs = " ";
popts->allow_repeat_ifs = TRUE;
}
else if (streq(argv[argi], "--ps")) {
check_arg_count(argv, argi, argc, 2);
popts->ops = sep_from_arg(argv[argi+1], argv[0]);
popts->ips = sep_from_arg(argv[argi+1], argv[0]);
argi++;
}
else if (streq(argv[argi], "--ips")) {
check_arg_count(argv, argi, argc, 2);
popts->ips = sep_from_arg(argv[argi+1], argv[0]);
argi++;
}
else if (streq(argv[argi], "--ops")) {
check_arg_count(argv, argi, argc, 2);
popts->ops = sep_from_arg(argv[argi+1], argv[0]);
argi++;
}
else if (streq(argv[argi], "--csv")) { popts->ifile_fmt = popts->ofile_fmt = "csv"; }
else if (streq(argv[argi], "--icsv")) { popts->ifile_fmt = "csv"; }
else if (streq(argv[argi], "--ocsv")) { popts->ofile_fmt = "csv"; }
else if (streq(argv[argi], "--csvlite")) { popts->ifile_fmt = popts->ofile_fmt = "csvlite"; }
else if (streq(argv[argi], "--icsvlite")) { popts->ifile_fmt = "csvlite"; }
else if (streq(argv[argi], "--ocsvlite")) { popts->ofile_fmt = "csvlite"; }
else if (streq(argv[argi], "--dkvp")) { popts->ifile_fmt = popts->ofile_fmt = "dkvp"; }
else if (streq(argv[argi], "--idkvp")) { popts->ifile_fmt = "dkvp"; }
else if (streq(argv[argi], "--odkvp")) { popts->ofile_fmt = "dkvp"; }
else if (streq(argv[argi], "--nidx")) { popts->ifile_fmt = popts->ofile_fmt = "nidx"; }
else if (streq(argv[argi], "--inidx")) { popts->ifile_fmt = "nidx"; }
else if (streq(argv[argi], "--onidx")) { popts->ofile_fmt = "nidx"; }
else if (streq(argv[argi], "--xtab")) { popts->ifile_fmt = popts->ofile_fmt = "xtab"; }
else if (streq(argv[argi], "--ixtab")) { popts->ifile_fmt = "xtab"; }
else if (streq(argv[argi], "--oxtab")) { popts->ofile_fmt = "xtab"; }
else if (streq(argv[argi], "--ipprint")) {
popts->ifile_fmt = "csvlite";
popts->ifs = " ";
popts->allow_repeat_ifs = TRUE;
}
else if (streq(argv[argi], "--opprint")) {
popts->ofile_fmt = "pprint";
}
else if (streq(argv[argi], "--pprint")) {
popts->ifile_fmt = "csvlite";
popts->ifs = " ";
popts->allow_repeat_ifs = TRUE;
popts->ofile_fmt = "pprint";
}
else if (streq(argv[argi], "--right")) {
left_align_pprint = FALSE;
}
else if (streq(argv[argi], "--ofmt")) {
check_arg_count(argv, argi, argc, 2);
popts->ofile_fmt = argv[argi+1];
argi++;
}
else if (streq(argv[argi], "--quote-all")) { popts->oquoting = QUOTE_ALL; }
else if (streq(argv[argi], "--quote-none")) { popts->oquoting = QUOTE_NONE; }
else if (streq(argv[argi], "--quote-minimal")) { popts->oquoting = QUOTE_MINIMAL; }
else if (streq(argv[argi], "--quote-numeric")) { popts->oquoting = QUOTE_NUMERIC; }
// xxx put into online help.
else if (streq(argv[argi], "--mmap")) {
popts->use_mmap_for_read = TRUE;
}
else if (streq(argv[argi], "--no-mmap")) {
popts->use_mmap_for_read = FALSE;
}
else if (streq(argv[argi], "--seed")) {
check_arg_count(argv, argi, argc, 2);
if (sscanf(argv[argi+1], "0x%x", &rand_seed) == 1) {
have_rand_seed = TRUE;
} else if (sscanf(argv[argi+1], "%u", &rand_seed) == 1) {
have_rand_seed = TRUE;
} else {
main_usage(argv[0], 1);
}
argi++;
}
else
nusage(argv[0], argv[argi]);
}
lhmss_t* default_rses = get_default_rses();
lhmss_t* default_fses = get_default_fses();
lhmss_t* default_pses = get_default_pses();
lhmsi_t* default_repeat_ifses = get_default_repeat_ifses();
lhmsi_t* default_repeat_ipses = get_default_repeat_ipses();
if (popts->irs == NULL)
popts->irs = lhmss_get(default_rses, popts->ifile_fmt);
if (popts->ifs == NULL)
popts->ifs = lhmss_get(default_fses, popts->ifile_fmt);
if (popts->ips == NULL)
popts->ips = lhmss_get(default_pses, popts->ifile_fmt);
if (popts->allow_repeat_ifs == NEITHER_TRUE_NOR_FALSE)
popts->allow_repeat_ifs = lhmsi_get(default_repeat_ifses, popts->ifile_fmt);
if (popts->allow_repeat_ips == NEITHER_TRUE_NOR_FALSE)
popts->allow_repeat_ips = lhmsi_get(default_repeat_ipses, popts->ifile_fmt);
if (popts->ors == NULL)
popts->ors = lhmss_get(default_rses, popts->ofile_fmt);
if (popts->ofs == NULL)
popts->ofs = lhmss_get(default_fses, popts->ofile_fmt);
if (popts->ops == NULL)
popts->ops = lhmss_get(default_pses, popts->ofile_fmt);
if (popts->irs == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->ifs == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->ips == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->allow_repeat_ifs == NEITHER_TRUE_NOR_FALSE) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->allow_repeat_ips == NEITHER_TRUE_NOR_FALSE) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->ors == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->ofs == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (popts->ops == NULL) {
fprintf(stderr, "%s: internal coding error detected in file %s at line %d.\n", argv[0], __FILE__, __LINE__);
exit(1);
}
if (streq(popts->ofile_fmt, "pprint") && strlen(popts->ofs) != 1) {
fprintf(stderr, "%s: OFS for PPRINT format must be single-character; got \"%s\".\n",
argv[0], popts->ofs);
return NULL;
}
if (streq(popts->ofile_fmt, "dkvp"))
popts->plrec_writer = lrec_writer_dkvp_alloc(popts->ors, popts->ofs, popts->ops);
else if (streq(popts->ofile_fmt, "csv"))
popts->plrec_writer = lrec_writer_csv_alloc(popts->ors, popts->ofs, popts->oquoting);
else if (streq(popts->ofile_fmt, "csvlite"))
popts->plrec_writer = lrec_writer_csvlite_alloc(popts->ors, popts->ofs);
else if (streq(popts->ofile_fmt, "nidx"))
popts->plrec_writer = lrec_writer_nidx_alloc(popts->ors, popts->ofs);
else if (streq(popts->ofile_fmt, "xtab"))
popts->plrec_writer = lrec_writer_xtab_alloc(popts->ofs, popts->ops);
else if (streq(popts->ofile_fmt, "pprint"))
popts->plrec_writer = lrec_writer_pprint_alloc(popts->ors, popts->ofs[0], left_align_pprint);
else {
main_usage(argv[0], 1);
}
if ((argc - argi) < 1) {
main_usage(argv[0], 1);
}
popts->pmapper_list = sllv_alloc();
while (TRUE) {
check_arg_count(argv, argi, argc, 1);
char* verb = argv[argi];
mapper_setup_t* pmapper_setup = look_up_mapper_setup(verb);
if (pmapper_setup == NULL) {
fprintf(stderr, "%s: verb \"%s\" not found. Please use \"%s --help\" for a list.\n",
argv[0], verb, argv[0]);
exit(1);
}
if ((argc - argi) >= 2) {
if (streq(argv[argi+1], "-h") || streq(argv[argi+1], "--help")) {
pmapper_setup->pusage_func(argv[0], verb);
exit(0);
}
}
// It's up to the parse func to print its usage on CLI-parse failure.
mapper_t* pmapper = pmapper_setup->pparse_func(&argi, argc, argv);
if (pmapper == NULL) {
exit(1);
}
sllv_add(popts->pmapper_list, pmapper);
// xxx cmt
if (argi >= argc || !streq(argv[argi], "then"))
break;
argi++;
}
popts->filenames = &argv[argi];
// No filenames means read from standard input, and standard input cannot be mmapped.
if (argi == argc)
popts->use_mmap_for_read = FALSE;
popts->plrec_reader = lrec_reader_alloc(popts->ifile_fmt, popts->use_mmap_for_read,
popts->irs, popts->ifs, popts->allow_repeat_ifs, popts->ips, popts->allow_repeat_ips);
if (popts->plrec_reader == NULL)
main_usage(argv[0], 1);
if (have_rand_seed) {
mtrand_init(rand_seed);
} else {
mtrand_init_default();
}
return popts;
}
// ----------------------------------------------------------------
// xxx cmt stdout/err & 0/1
static void main_usage(char* argv0, int exit_code) {
FILE* o = exit_code == 0 ? stdout : stderr;
fprintf(o, "Usage: %s [I/O options] {verb} [verb-dependent options ...] {file names}\n", argv0);
fprintf(o, "Verbs:\n");
char* leader = " ";
char* separator = " ";
int leaderlen = strlen(leader);
int separatorlen = strlen(separator);
int linelen = leaderlen;
int j = 0;
for (int i = 0; i < mapper_lookup_table_length; i++) {
char* verb = mapper_lookup_table[i]->verb;
int verblen = strlen(verb);
linelen += separatorlen + verblen;
if (linelen >= 80) {
fprintf(o, "\n");
linelen = leaderlen + separatorlen + verblen;
j = 0;
}
if (j == 0)
fprintf(o, "%s", leader);
fprintf(o, "%s%s", separator, verb);
j++;
}
fprintf(o, "\n");
fprintf(o, "Example: %s --csv --rs lf --fs tab cut -f hostname,uptime file1.csv file2.csv\n", argv0);
fprintf(o, "Please use \"%s -h\" or \"%s --help\" to show this message.\n", argv0, argv0);
fprintf(o, "Please use \"%s --version\" to show the software version.\n", argv0);
fprintf(o, "Please use \"%s {verb name} --help\" for verb-specific help.\n", argv0);
fprintf(o, "Please use \"%s --help-all-verbs\" for help on all verbs.\n", argv0);
fprintf(o, "\n");
lrec_evaluator_list_functions(o);
fprintf(o, "Please use \"%s --help-function {function name}\" for function-specific help.\n", argv0);
fprintf(o, "Please use \"%s --help-all-functions\" or \"%s -f\" for help on all functions.\n", argv0, argv0);
fprintf(o, "\n");
fprintf(o, "Data-format options, for input, output, or both:\n");
fprintf(o, " --dkvp --idkvp --odkvp Delimited key-value pairs, e.g \"a=1,b=2\"\n");
fprintf(o, " (default)\n");
fprintf(o, " --nidx --inidx --onidx Implicitly-integer-indexed fields\n");
fprintf(o, " (Unix-toolkit style)\n");
fprintf(o, " --csv --icsv --ocsv Comma-separated value (or tab-separated\n");
fprintf(o, " with --fs tab, etc.)\n");
fprintf(o, " --pprint --ipprint --opprint --right Pretty-printed tabular (produces no\n");
fprintf(o, " output until all input is in)\n");
fprintf(o, " --xtab --ixtab --oxtab Pretty-printed vertical-tabular\n");
fprintf(o, " -p is a keystroke-saver for --nidx --fs space --repifs\n");
fprintf(o, "Separator options, for input, output, or both:\n");
fprintf(o, " --rs --irs --ors Record separators, e.g. 'lf' or '\\r\\n'\n");
fprintf(o, " --fs --ifs --ofs --repifs Field separators, e.g. comma\n");
fprintf(o, " --ps --ips --ops Pair separators, e.g. equals sign\n");
fprintf(o, " Notes:\n");
fprintf(o, " * IPS/OPS are only used for DKVP and XTAB formats, since only in these formats\n");
fprintf(o, " do key-value pairs appear juxtaposed.\n");
fprintf(o, " * IRS/ORS are ignored for XTAB format. Nominally IFS and OFS are newlines;\n");
fprintf(o, " XTAB records are separated by two or more consecutive IFS/OFS -- i.e.\n");
fprintf(o, " a blank line.\n");
fprintf(o, " * OFS must be single-character for PPRINT format. This is because it is used\n");
fprintf(o, " with repetition for alignment; multi-character separators would make\n");
fprintf(o, " alignment impossible.\n");
fprintf(o, " * OPS may be multi-character for XTAB format, in which case alignment is\n");
fprintf(o, " disabled.\n");
fprintf(o, " * DKVP, NIDX, CSVLITE, PPRINT, and XTAB formats are intended to handle\n");
fprintf(o, " platform-native text data. In particular, this means LF line-terminators\n");
fprintf(o, " by default on Linux/OSX. You can use \"--dkvp --rs crlf\" for\n");
fprintf(o, " CRLF-terminated DKVP files, and so on.\n");
fprintf(o, " * CSV is intended to handle RFC-4180-compliant data. In particular, this means\n");
fprintf(o, " it uses CRLF line-terminators by default. You can use \"--csv --rs lf\" for\n");
fprintf(o, " Linux-native CSV files.\n");
fprintf(o, " * You can specify separators in any of the following ways, shown by example:\n");
fprintf(o, " - Type them out, quoting as necessary for shell escapes, e.g.\n");
fprintf(o, " \"--fs '|' --ips :\"\n");
fprintf(o, " - C-style escape sequences, e.g. \"--rs '\\r\\n' --fs '\\t'\".\n");
fprintf(o, " - To avoid backslashing, you can use any of the following names:\n");
fprintf(o, " ");
lhmss_t* pmap = get_desc_to_chars_map();
for (lhmsse_t* pe = pmap->phead; pe != NULL; pe = pe->pnext) {
fprintf(o, " %s", pe->key);
}
fprintf(o, "\n");
fprintf(o, " * Default separators by format:\n");
fprintf(o, " %-12s %-8s %-8s %s\n", "File format", "RS", "FS", "PS");
lhmss_t* default_rses = get_default_rses();
lhmss_t* default_fses = get_default_fses();
lhmss_t* default_pses = get_default_pses();
for (lhmsse_t* pe = default_rses->phead; pe != NULL; pe = pe->pnext) {
char* filefmt = pe->key;
char* rs = pe->value;
char* fs = lhmss_get(default_fses, filefmt);
char* ps = lhmss_get(default_pses, filefmt);
fprintf(o, " %-12s %-8s %-8s %s\n", filefmt, rebackslash(rs), rebackslash(fs), rebackslash(ps));
}
fprintf(o, "Double-quoting for CSV output:\n");
fprintf(o, " --quote-all Wrap all fields in double quotes\n");
fprintf(o, " --quote-none Do not wrap any fields in double quotes, even if they have \n");
fprintf(o, " OFS or ORS in them\n");
fprintf(o, " --quote-minimal Wrap fields in double quotes only if they have OFS or ORS\n");
fprintf(o, " in them (default)\n");
fprintf(o, " --quote-numeric Wrap fields in double quotes only if they have numbers\n");
fprintf(o, " in them\n");
fprintf(o, "Numerical formatting:\n");
fprintf(o, " --ofmt {format} E.g. %%.18lf, %%.0lf. Please use sprintf-style codes for\n");
fprintf(o, " double-precision. Applies to verbs which compute new\n");
fprintf(o, " values, e.g. put, stats1, stats2. See also the fmtnum\n");
fprintf(o, " function within mlr put (mlr --help-all-functions).\n");
fprintf(o, "Other options:\n");
fprintf(o, " --seed {n} with n of the form 12345678 or 0xcafefeed. For put/filter urand().\n");
fprintf(o, "Output of one verb may be chained as input to another using \"then\", e.g.\n");
fprintf(o, " %s stats1 -a min,mean,max -f flag,u,v -g color then sort -f color\n", argv0);
fprintf(o, "For more information please see http://johnkerl.org/miller/doc and/or\n");
fprintf(o, "http://github.com/johnkerl/miller.");
#ifdef HAVE_CONFIG_H
fprintf(o, " This is Miller version >= %s.\n", PACKAGE_VERSION);
#else
fprintf(o, " This is Miller version >= %s.\n", MLR_VERSION);
#endif // HAVE_CONFIG_H
exit(exit_code);
}
