static int inspect_top_left (xmlNodePtr p, int *obscol)
{
char *content = NULL;
int err, vtype = 0;
err = node_get_vtype_and_content(p, &vtype, &content);
if (!err) {
if (vtype == VALUE_EMPTY) {
*obscol = 1;
} else if (vtype == VALUE_STRING) {
if (import_obs_label(content)) {
*obscol = 1;
}
}
}
free(content);
return err;
}
static void xlsx_check_top_left (xlsx_info *xinfo, int r, int c,
int stringcell, const char *s,
double x)
{
if (r == xinfo->yoffset + 1 && c == xinfo->xoffset + 1) {
/* We're in the top left cell of the reading area:
this could be blank, or could hold the first
varname, could hold "obs" or similar, or could
be the first numerical value.
*/
#if XDEBUG
fprintf(stderr, "xlsx_check_top_left: r=%d, c=%d, x=%g, stringcell=%d, "
"s='%s'\n", r, c, x, stringcell, s);
#endif
if (!na(x)) {
/* got a valid numerical value: that means we don't
have variable names on the top row */
xinfo->flags |= BOOK_AUTO_VARNAMES;
} else if (stringcell && import_obs_label(s)) {
/* blank or "obs" or similar */
xinfo->flags |= BOOK_OBS_LABELS;
xinfo->obscol = c;
}
if (!na(x) || stringcell) {
/* record the fact that the top-left corner is not empty */
xinfo->flags &= ~BOOK_TOP_LEFT_EMPTY;
}
} else if (r == xinfo->yoffset + 1 && c == xinfo->xoffset + 2) {
/* first row, second column */
if (!na(x)) {
/* got a number, not a varname */
xinfo->flags |= BOOK_AUTO_VARNAMES;
} else {
xinfo->namerow = r;
}
}
}
int gnumeric_get_data (const char *fname, int *list, char *sheetname,
DATASET *dset, gretlopt opt, PRN *prn)
{
int gui = (opt & OPT_G);
wbook gbook;
wbook *book = &gbook;
wsheet gsheet;
wsheet *sheet = &gsheet;
int sheetnum = -1;
DATASET *newset;
int err = 0;
newset = datainfo_new();
if (newset == NULL) {
pputs(prn, _("Out of memory\n"));
return 1;
}
wsheet_init(sheet);
gretl_push_c_numeric_locale();
if (wbook_get_info(fname, list, sheetname, book, prn)) {
pputs(prn, _("Failed to get workbook info"));
err = 1;
goto getout;
}
wbook_print_info(book);
if (book->nsheets == 0) {
pputs(prn, _("No worksheets found"));
err = 1;
goto getout;
}
if (gui) {
if (book->nsheets > 1) {
wsheet_menu(book, 1);
sheetnum = book->selected;
} else {
wsheet_menu(book, 0);
sheetnum = 0;
}
} else {
err = wbook_check_params(book);
if (err) {
gretl_errmsg_set(_("Invalid argument for worksheet import"));
} else if (book->selected >= 0) {
sheetnum = book->selected;
} else {
sheetnum = 0;
}
}
if (book->selected == -1) {
/* canceled */
err = -1;
}
if (!err && sheetnum >= 0) {
fprintf(stderr, "Getting data...\n");
if (wsheet_setup(sheet, book, sheetnum)) {
pputs(prn, _("error in wsheet_setup()"));
err = 1;
} else {
err = wsheet_get_data(fname, sheet, prn);
if (!err) {
wsheet_print_info(sheet);
book->flags |= sheet->flags;
}
}
}
if (err) {
goto getout;
} else {
int r0 = 1;
int i, j, t;
int ts_markers = 0;
int merge = (dset->Z != NULL);
char **ts_S = NULL;
int blank_cols = 0;
int missvals = 0;
int pd = 0;
if (sheet->text_cols > 0) {
book_set_obs_labels(book);
}
if (sheet->colheads == 0) {
book_set_auto_varnames(book);
r0 = 0;
}
if (book_numeric_dates(book)) {
puts("found calendar dates in first imported column");
} else if (sheet->text_cols > 0) {
puts("found label strings in first imported column");
} else {
puts("check for label strings in first imported column: not found");
}
newset->n = sheet->maxrow - sheet->row_offset;
if (!sheet->colheads) {
pputs(prn, _("it seems there are no variable names\n"));
newset->n += 1;
}
if (book_numeric_dates(book) ||
(sheet->colheads > 0 && import_obs_label(sheet->label[0]))) {
pd = importer_dates_check(sheet->label + r0, &book->flags,
newset, prn, &err);
if (pd > 0) {
/* got time-series info from dates/labels */
sheet_time_series_setup(sheet, book, newset, pd);
ts_markers = newset->markers;
ts_S = newset->S;
} else if (!book_numeric_dates(book)) {
if (labels_numeric(sheet->label, newset->n)) {
sheet->text_cols = 0;
book_unset_obs_labels(book);
}
}
}
newset->v = sheet->maxcol + 2 - sheet->col_offset - sheet->text_cols;
fprintf(stderr, "newset->v = %d, newset->n = %d\n",
newset->v, newset->n);
/* create import dataset */
err = worksheet_start_dataset(newset);
if (err) {
goto getout;
}
if (book_time_series(book)) {
newset->markers = ts_markers;
newset->S = ts_S;
} else {
dataset_obs_info_default(newset);
}
j = 1;
for (i=1; i<newset->v; i++) {
int s = (sheet->colheads)? 1 : 0;
int k = i - 1 + sheet->text_cols;
double zkt;
if (column_is_blank(sheet, k, newset->n)) {
blank_cols++;
continue;
}
if (sheet->colheads && *sheet->varname[k] != '\0') {
strcpy(newset->varname[j], sheet->varname[k]);
} else {
sprintf(newset->varname[j], "v%d", j);
}
for (t=0; t<newset->n; t++) {
zkt = sheet->Z[k][s++];
if (zkt == -999 || zkt == -9999) {
newset->Z[j][t] = NADBL;
} else {
newset->Z[j][t] = zkt;
}
if (na(newset->Z[j][t])) {
missvals = 1;
}
}
j++;
}
if (blank_cols > 0) {
fprintf(stderr, "Dropping %d apparently blank column(s)\n",
blank_cols);
dataset_drop_last_variables(newset, blank_cols);
}
if (missvals) {
pputs(prn, _("Warning: there were missing values\n"));
}
if (fix_varname_duplicates(newset)) {
pputs(prn, _("warning: some variable names were duplicated\n"));
}
if (book_obs_labels(book) && wsheet_labels_complete(sheet)) {
int offset = (sheet->colheads)? 1 : 0;
dataset_allocate_obs_markers(newset);
if (newset->S != NULL) {
for (t=0; t<newset->n; t++) {
strcpy(newset->S[t], sheet->label[t+offset]);
}
}
}
if (book->flags & BOOK_DATA_REVERSED) {
reverse_data(newset, prn);
}
if (!err && !dataset_is_time_series(newset) && newset->S != NULL) {
/* we didn't time series info above, but it's possible
the observation strings carry such info
*/
import_ts_check(newset);
}
err = merge_or_replace_data(dset, &newset, opt, prn);
if (!err && !merge) {
dataset_add_import_info(dset, fname, GRETL_GNUMERIC);
}
if (!err && gui) {
wbook_record_params(book, list);
}
}
getout:
wbook_free(book);
wsheet_free(sheet);
gretl_pop_c_numeric_locale();
if (err && newset != NULL) {
destroy_dataset(newset);
}
return err;
}
static int wsheet_parse_cells (xmlNodePtr node, wsheet *sheet, PRN *prn)
{
xmlNodePtr p = node->xmlChildrenNode;
char *tmp;
double x;
int vtype = 0;
int gotlabels = 0;
int cols, rows;
int i, t, r, c;
int err = 0;
cols = sheet->maxcol + 1 - sheet->col_offset;
rows = sheet->maxrow + 1 - sheet->row_offset;
if (rows < 1) {
pputs(prn, _("Starting row is out of bounds.\n"));
return 1;
}
if (cols < 1) {
pputs(prn, _("Starting column is out of bounds.\n"));
return 1;
}
if (wsheet_allocate(sheet, cols, rows)) {
return 1;
}
sheet->colheads = 0;
while (p != NULL && !err) {
if (!xmlStrcmp(p->name, (XUC) "Cell")) {
x = NADBL;
c = r = 0;
i = t = -1;
/* what column are we in? */
tmp = (char *) xmlGetProp(p, (XUC) "Col");
if (tmp) {
c = atoi(tmp);
i = c - sheet->col_offset;
free(tmp);
}
/* what row are we on? */
tmp = (char *) xmlGetProp(p, (XUC) "Row");
if (tmp) {
r = atoi(tmp);
t = r - sheet->row_offset;
free(tmp);
}
if (i < 0 || t < 0) {
/* we're not in the requested reading area */
p = p->next;
continue;
}
/* check that we have a value of some type available */
tmp = (char *) xmlGetProp(p, (XUC) "ValueType");
if (tmp) {
vtype = atoi(tmp);
free(tmp);
} else {
/* a formula perhaps? */
pprintf(prn, _("Couldn't get value for col %d, row %d.\n"
"Maybe there's a formula in the sheet?"),
c, r);
err = 1;
break;
}
/* get and process the actual cell content */
tmp = (char *) xmlNodeGetContent(p);
if (tmp != NULL) {
if (VTYPE_IS_NUMERIC(vtype) || vtype == VALUE_STRING) {
if (i == 0) {
/* first column: write content to labels */
gretl_utf8_strncat_trim(sheet->label[t], tmp, OBSLEN - 1);
}
}
if (i == 0 && t == 1 && VTYPE_IS_NUMERIC(vtype)) {
char *fmt = (char *) xmlGetProp(p, (XUC) "ValueFormat");
if (fmt) {
check_for_date_format(sheet, fmt);
free(fmt);
}
}
if (VTYPE_IS_NUMERIC(vtype)) {
x = atof(tmp);
sheet->Z[i][t] = x;
} else if (i > 0 && stray_numeric(vtype, tmp, &x)) {
sheet->Z[i][t] = x;
} else if (vtype == VALUE_STRING) {
if (t == 0) {
/* first row: look for varnames */
strncat(sheet->varname[i], tmp, VNAMELEN - 1);
sheet->colheads += 1;
if (i == 0 && import_obs_label(tmp)) {
; /* keep going */
} else {
err = check_imported_varname(sheet->varname[i],
r, c, prn);
}
} else if (i == 0) {
/* first column, not first row */
if (!gotlabels) {
gotlabels = 1;
}
sheet->text_cols = 1;
} else {
pprintf(prn, _("Expected numeric data, found string:\n"
"'%s' at row %d, column %d\n"),
tmp, r, c);
err = 1;
}
}
free(tmp);
}
}
p = p->next;
}
if (gotlabels && sheet->colheads == 1) {
/* rough notion here: if there's only one heading, it's
probably not really a variable name, but rather
a first observation label
*/
sheet->colheads = 0;
}
return err;
}
