A bug with "Replace all" functionality in Word

[HieuMark]

I found out that there is an issue with body_replace_all_text() when used with "Replace all" in Words.

Here is my experiment in case anyone wants to try this themself:

1. Create an empty .docx file, add "{{name}}" in it.
2. Use the "Find and replace" functionality in Word to replace "{{" -> "(", and "}}" -> ")" (it will turn into "(name)"), then Save the .docx file (don't change it manually).
3. Run the following code:

test_doc = officer::read_docx("../test_doc.docx")
test_sum = officer::docx_summary(test_doc)

test_doc = officer::body_replace_all_text(
  test_doc,
  old_value = "(name)",
  new_value = "my new value",
  fixed = TRUE
)
output_sum = officer::docx_summary(test_doc)

print(test_doc, target = "../test_output.docx")

The output file remains with "(name)", because body_replace_all_text() cannot find "(name)" that was replaced using "Find and replace". It would be nice if anyone has an idea why this is the case.

TLDR: body_replace_all_text() and Word's "Find and replace" do not work with each other.

[davidgohel]

See https://davidgohel.github.io/officer/reference/body_replace_all_text.html section Chunking of text:

Note that the behind-the-scenes representation of text in a Word document is frequently not what you might expect! Sometimes a paragraph of text is broken up (or "chunked") into several "runs," as a result of style changes, pauses in text entry, later revisions and edits, etc. If you have not styled the text, and have entered it in an "all-at-once" fashion, e.g. by pasting it or by outputing it programmatically into your Word document, then this will likely not be a problem. If you are working with a manually-edited document, however, this can lead to unexpected failures to find text.
You can use the officer function docx_show_chunk() to show how the paragraph of text at the current cursor has been chunked into runs, and what text is in each chunk. This can help troubleshoot unexpected failures to find text.