augustoasilva commented on a change in pull request #11551:
URL: https://github.com/apache/arrow/pull/11551#discussion_r760060674
##########
File path: cpp/src/gandiva/gdv_function_stubs.cc
##########
@@ -832,6 +850,201 @@ const char* gdv_fn_initcap_utf8(int64_t context, const
char* data, int32_t data_
*out_len = out_idx;
return out;
}
+
+GANDIVA_EXPORT
+const char* gdv_mask_first_n_utf8_int32(int64_t context, const char* data,
+ int32_t data_len, int32_t n_to_mask,
+ int32_t* out_len) {
+ if (data_len <= 0) {
+ *out_len = 0;
+ return nullptr;
+ }
+
+ if (n_to_mask > data_len) {
+ n_to_mask = data_len;
+ }
+
+ *out_len = data_len;
+
+ if (n_to_mask <= 0) {
+ return data;
+ }
+
+ char* out = reinterpret_cast<char*>(gdv_fn_context_arena_malloc(context,
*out_len));
+ if (out == nullptr) {
+ gdv_fn_context_set_error_msg(context, "Could not allocate memory for
output string");
+ *out_len = 0;
+ return nullptr;
+ }
+
+ int bytes_masked;
+ for (bytes_masked = 0; bytes_masked < n_to_mask; bytes_masked++) {
+ unsigned char char_single_byte = data[bytes_masked];
+ if (char_single_byte > 127) {
+ // found a multi-byte utf-8 char
+ break;
+ }
+ out[bytes_masked] = mask_array[char_single_byte];
+ }
+
+ int chars_masked = bytes_masked;
+ int out_idx = bytes_masked;
+
+ // Handle multibyte utf8 characters
+ utf8proc_int32_t utf8_char;
+ while ((chars_masked < n_to_mask) && (bytes_masked < data_len)) {
+ auto char_len =
+ utf8proc_iterate(reinterpret_cast<const utf8proc_uint8_t*>(data +
bytes_masked),
+ data_len, &utf8_char);
+
+ if (char_len < 0) {
+ gdv_fn_context_set_error_msg(context, utf8proc_errmsg(char_len));
+ *out_len = 0;
+ return nullptr;
+ }
+
+ switch (utf8proc_category(utf8_char)) {
+ case 1:
+ out[out_idx] = 'X';
+ out_idx++;
+ break;
+ case 2:
+ out[out_idx] = 'x';
+ out_idx++;
+ break;
+ case 9:
+ out[out_idx] = 'n';
+ out_idx++;
+ break;
+ case 10:
+ out[out_idx] = 'n';
+ out_idx++;
+ break;
+ default:
+ memcpy(out + out_idx, data + bytes_masked, char_len);
+ out_idx += static_cast<int>(char_len);
+ break;
+ }
+ bytes_masked += static_cast<int>(char_len);
+ chars_masked++;
+ }
+
+ // Correct the out_len after masking multibyte characters with single byte
characters
+ *out_len = *out_len - (bytes_masked - out_idx);
+
+ if (bytes_masked < data_len) {
+ memcpy(out + out_idx, data + bytes_masked, data_len - bytes_masked);
+ }
+
+ return out;
+}
+
+GANDIVA_EXPORT
+const char* gdv_mask_last_n_utf8_int32(int64_t context, const char* data,
+ int32_t data_len, int32_t n_to_mask,
+ int32_t* out_len) {
+ if (data_len <= 0) {
+ *out_len = 0;
+ return nullptr;
+ }
+
+ if (n_to_mask > data_len) {
+ n_to_mask = data_len;
+ }
+
+ *out_len = data_len;
+
+ if (n_to_mask <= 0) {
+ return data;
+ }
+
+ char* out = reinterpret_cast<char*>(gdv_fn_context_arena_malloc(context,
*out_len));
+ if (out == nullptr) {
+ gdv_fn_context_set_error_msg(context, "Could not allocate memory for
output string");
+ *out_len = 0;
+ return nullptr;
+ }
+
+ bool has_multi_byte = false;
+ for (int i = 0; i < data_len; i++) {
+ unsigned char char_single_byte = data[i];
+ if (char_single_byte > 127) {
+ // found a multi-byte utf-8 char
+ has_multi_byte = true;
+ break;
+ }
+ }
+
+ if (!has_multi_byte) {
+ int start_idx = data_len - n_to_mask;
+ for (int i = start_idx; i < data_len; ++i) {
+ unsigned char char_single_byte = data[i];
+ out[i] = mask_array[char_single_byte];
+ }
+ memcpy(out, data, start_idx);
+ *out_len = data_len;
+ return out;
+ }
+
+ utf8proc_int32_t utf8_char_buffer;
+ int num_of_chars = static_cast<int>(
+ utf8proc_decompose(reinterpret_cast<const utf8proc_uint8_t*>(data),
data_len,
Review comment:
It loops through the entire string looking for chars (valid codepoints)
acording to the option that is passed. Differently from the utf8proc_iterate()
that loops through a sequence of bytes to find the first valid char (valid code
point).
It does not, the second data_len is passed as the size of the buffer to be
used, for the tests I used data_len and 4 (as the utf8 chars can have only up
to 4 bytes) And if forgot to change this to 4 before pushing it to the repo.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: github-unsubscr...@arrow.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
