Using Wildcard Search and Replace to switch from manual numbering to automatic numbering

This is again in response to a question from the Word-PC e-mail list.

You have a document where the numbering has been entered manually, and you would like to convert this into a document which uses the different heading styles with associated numbering.

This is a slightly modified version of Example X (that’s ten, by the way) from Application 10.3 (Practical wildcard searches) in my book Doing your dissertation with Microsoft Word.

Students have manually numbered and formatted their thesis before learning about the principles taught in this book which allows Word to do their numbering for them. They want to start using this system (e.g., so as to still be able to use Word to create their table of contents), but cannot simply apply the heading styles, as all the old numbers still have to be deleted manually, and in any case it would be great if Word could actually find all their headings and update them to the styles automatically, without the students having to apply the style to each heading.

The secret is to start with the lowest heading level, hierarchically speaking. Thus, if your numbering is like this, then level 4 is your lowest hierarchical level:


Numbering example


Chapter 1






Since the principle is the same for each level, I will only do levels three and higher in this example.

Also note that this assumes that the students very consistently did their numbering—they missed no separating periods, made no formatting mistakes, etc. Every time one of those mistakes is made, it means that the search and replace might miss something. In short, given the human propensity for error, it is best to first make a backup of the document (use my Word uTIlities to make that easy), and then to do a thorough check once the search and replace process has been completed.

For heading level three, the search specification is “([1-6].[1-9].[1-9] )(*^13)”—obviously without the quotes, here and further—which assumes that the students did not ever use more than nine third level subheadings to a single second level heading, the same for second to first level headings, and that the student has six chapters in the thesis. When in doubt, go big and specify it as follows: “([1-9].[0-9]{1,2}.[0-9]{1,2} )(*^13)” which allows for up to a single digit first level, a two digit second level, and a two digit third level (i.e., 1.1.1 through 9.99.99). Very important here is to note the space before the first right parenthesis.

To explain what this is doing, the search specification consists of two groups (as defined by the two sets of parentheses): The first group consists of the number, which is, in the first example, a digit between 1 and 6 for the chapter, followed by a period, a digit between 1 and 9 for the second level heading, followed by a period, and a digit between 1 and 9 for the third level heading, followed by a space. The second group is any text (in this case, the heading text) up to a paragraph mark. The replace with criteria is simply \2 with the formatting specification of Style: Heading 3 (Click on Format, Style in the Find and Replace dialog, and choose Heading 3). This (\2) puts only group 2 (the text and the paragraph mark) back in the replace action. The style is what adds the numbering.

Thus the number is stripped away, and the style applied, all at the same time.

In addition, the formatting for the Find What box should be set to the formatting used for the third level heading. This is actually quite necessary, as without that, the search and replace will also find cross-references in the text to those heading numbers, which is not what you want to do!

So it would look something like this:

The process can then be repeated, with ([1-6].[1-9] )(*^13) for the second level heading, and exactly the same replace with specification, except that now the formatting is set to Style: Heading 2, and of course the appropriate formatting for the Find what box too.

For the Chapter titles, again the appropriate formatting for the Find what box is set, and then the specification is (Chapter [1-6] )(*^13) with exactly the same replace with criteria and with the formatting specification of Style: Heading 1. In this way, an entire manually numbered thesis can be transformed into an automatically numbered thesis in a matter of minutes.

Since you have made a backup of the document, give it a bash, thrash it out, and see if you can get it working. I have actually used this technique on a number of theses already, and while I do concede that it is complex and heady, I can also attest that it is amazingly effective, giving great results and saving hours of manual work. I have sometimes just done a few Find Next operations until I am certain that it is finding what it needs to be finding, and is not finding the cross references, and then, when I feel bold enough, I hit Replace All (easy enough to do if you have a backup!).


Separate table (or figure) numbering for appendices

A question recently came up in the Word-PC e-mail list which is quite typical of something a student doing a thesis or dissertation might come across. I will also modify this post to that context. I was quick to give an answer, and only later thought through all the implications, so that my initial answer would have been less than satisfactory. Here is, what I believe, a more robust solution.

The requirement is as follows:

  1. A dissertation is using chapter-based numbering instead of sequential numbering for its tables (i.e., in Chapter 3, the tables would be 3-1, 3-2, etc.). Note of course, that the student can set the delimiter (here a hyphen) in the process of activating chapter based numbering (the dialog is shown below)—many students use a period.
  2. The dissertation also has appendices, which must be numbered with uppercase letters (Appendix A, Appendix B, etc.).
  3. Those appendices contain tables (or figures, or any other label created with the Insert Caption tool).
  4. Those tables need to be numbered in the chapter-based style, but now reflecting the Appendix number. Thus A-1, A-2, etc. for Appendix A, and so on.

Parts of these are very easy to do. It’s the combination that is a headache.

Requirement 1 is simple. Just use the Insert Caption tool to add the numbers, and set the table-based numbering. See p. 162–163 of Doing your dissertation with Microsoft Word if you have the book (if you don’t some screen caps below will help).

Requirement 2 is equally simple. I have created, in the two templates (Thesis 2010.dotx and Thesis up 2009.dotx) which can be downloaded from the examples page, a style called “Appendix Title” that uses normal upper case alphabetic numbering. This is what the style looks like, showing its definition:

Some pointers. When modifying the style, be sure to go to the paragraph dialog and set its outline level to Level 1. This will ensure that it should be included in your Table of Contents (mine is based on Heading 1, which gives the same end result). Using this style for your appendix headings gives them the desired numbering scheme, thus satisfying requirement 2.

Requirements 3 and 4 are dealt with together, as these are the more tricky ones.

Here are the problems:

When Word adds caption labels, and those caption labels use chapter-based numbering, Word actually adds two fields: One is a STYLEREF field which is set to automatically pull in the heading level 1. The second is a SEQ field, which is set to restart after each occurrence of heading level 1.

Furthermore, whenever you add a table anywhere in the dissertation, Word runs through all instances of the table, and resets all of them. If you thus make changes to the switches of these caption labels in the appendix, they can get nixed by adding any table anywhere else in the dissertation.

To get started, use the Insert Caption tool to insert the caption label for the first table in the first appendix. We will modify the fields it adds afterwards (or, if you are confident writing out fields by hand, you can just skip ahead).

Click on Insert Caption (Captions group, References ribbon). From the Caption dialog, choose Numbering… to open the Caption Numbering dialog, and include the chapter number, also choosing the delimiter (separator). No need to change the Chapter start with style setting here, since all that is allow are the nine built-in heading levels. We will modify this manually later:

Once the chapter-based numbering has been set, the Caption dialog should look more or less like this:

Clicking on OK let’s Word add the caption label, using two fields.

Let’s first look at those fields Word adds. This is what they look like before and after being selected and set to Show field codes (Shift + F9):

So. The field codes (the bits between the braces) and text Word adds, is:

Table { STYLEREF 1 \s }–{SEQ Table \* ARABIC \s 1}

We can’t just change the STYLEREF field from 1 (Word shorthand for the built-in style “Heading 1”) to “Appendix Title,” as Word will nix this each time it updates the caption labels when a table is added somewhere. We also cannot change the restart switch for the SEQ field to anything other than 1-9 (i.e., Word’s nine built-in heading styles).

So we need a slightly different approach.

The first thing is to keep Word from automatically changing these table captions. The second is to get it to restart with each new appendix. Both of these are accomplished by setting the first appendix table caption to:

Table { STYLEREF “Appendix Title” \s }–{SEQ AppendixTable \* ARABIC \r 1}

The change to the STYLEREF field will pull in the required appendix number (from my custom style–if you are using a different style, then you would have to include whatever that style name would be). But note that it is the change to the SEQ identifier (from Table to AppendixTable) that essentially creates a new sequence, and prevents Word from reverting the STYLEREF field from STYLEREF “Appendix Title” back to STYLEREF 1. The SEQ field switch \r 1 which has replaced the \s 1 switch tells Word to restart at 1 here. Note that the 1s in the two switches have vastly different meanings. The \s 1 means restart numbering if there is an instance of “Heading 1” style between me and the preceding similar caption. The 1 refers to Heading 1. The \r 1 means restart right here, using the number 1. Here the 1 refers to 1.

But still more needs to be done. For each subsequent table in each appendix, the caption must be:

Table { STYLEREF “Appendix Title” \s }–{SEQ AppendixTable \* ARABIC \n}

Here the \n switch just tells Word to continue the numbering sequence from the previous similar caption label.

So the first table caption of each appendix will be:

Table { STYLEREF “Appendix Title” \s }–{SEQ AppendixTable \* ARABIC \r 1}

And the table caption of all other tables in each appendix will be:

Table { STYLEREF “Appendix Title” \s }–{SEQ AppendixTable \* ARABIC \n}

Probably the best way to achieve this is to simply copy and paste the caption paragraph for each new table. Yes, it’s a little manual work, but it’s not that bad. If you’re work averse, you could always record a macro writing out the fields (remember to use Ctrl + F9 to add the field braces) and then run that macro each time. Remember to also set the paragraph style to Caption, thus ensuring consistent formatting for your chapter and appendix tables.

Unfortunately, the job is not quite done.

Because we changed the SEQ field from SEQ Table to SEQ AppendixTable, the tables in the appendices will not appear in the list of tables. So we need to fix that.

To get that done, insert (or re-insert) the list of tables.

From the Table of Figures dialog, select the Options button:

In the Table of Figures Options dialog, specify that the table of figures should be built from the Caption Style:

Once you click on OK, notice that back in the Table of Figures dialog, Word will have set the Caption label to (none). This won’t do, because that means that the table will include, tables, figures, and any other caption labels. Compare the Table of Figures dialog below to the one shown above:

Just reset this to Table, and click on OK:

The end result is a list of tables that includes both your chapter and appendix tables:

The great thing about this approach is that it is stable and it works. It will withstand the updating caused by adding new tables or deleting existing tables, updating all fields in the document, closing and re-opening the document, etc.

It does require a small amount of manual work, but I think it is well worth the effort.

Update (2014-03-11):
If you want to do this with multiple captions (e.g., Tables and Figures in your appendices), then a little more work is needed. Read about it here.


Numbered equations in Word

Microsoft boasts of its academic staff, but really, I often get the impression that no-one involved with Word has any idea of what academics do, need, or want. When the referencing tools in Word came out I got all excited, only to be severely disappointed (you can read about all my gripes in my book). Basically, after more than six years now (from Word 2007 to Word 2013), Microsoft is still more than two decades behind in the reference management field. Then again, is it fair to expect Word to do what is actually the domain of a whole difference class of software? Probably not, but then why did Microsoft even try?

One thing that Word should be able to do, and that you will find people regularly complaining about, is to neatly number equations. It seems that waaaaaaay back (I can’t even recall how far back, but I am certain that it was pre-Word 95), someone at Microsoft thought that being able to add a caption titled “Equation” with the Insert Caption tool was all that any academic would ever want (they probably have never worked through a statistics textbook).

The generally touted workaround (it remains a workaround until Microsoft actually addresses the problem, which seems as likely as Microsoft ever building true 3D capabilities into Excel) is this: Add an unbordered table consisting of one row and three columns, and then add the equation in the middle column, and the number in the right column. See, for example:

Here’s my suggested method. I admit that it still requires more manual labour than I feel should be necessary, but I do believe it presents a more robust solution than that presented by Microsoft’s employees. The first step is to add two blank paragraphs, position yourself in the first of these, and then add the equation. Then position yourself to the right of the equation (i.e., just before the first of the two paragraph marks), and press Ctrl + Alt + Enter to add a Style Separator. Now add a tab, and then insert an Equation caption (remembering to set the numbering so as to include the chapter number, and the caption itself so as to Exclude the label from the caption).

Adding an equation caption

The convention, again, is for the equations to be enclosed in parentheses, which can now be added.

This will leave your equations, properly centred on the page, with auto-numbering equation captions properly aligned, that can easily be used to generate a list of equations using the Insert Table of Figures tool.
Admittedly, this may not be an ideal list of equations, as the fact that the actual caption “Equation” is excluded from the caption, means that the list will not contain entries like “Equation 2.1                p.13,” but rather “(2.1)            p.13.”
However, using some bookmark-modification techniques, the parentheses can easily be removed from the reference, and with some creativity, the remaining list can be formatted quite effectively (this need only be done once). Also, very often the idea behind numbering is not to generate a list of the actual equations, but rather to have an auto-numbering list that can be cross-referenced. Because the Caption tool uses the SEQ field, it remains ideal for this purpose. This list will show the equations with the right numbers, in the right order, with the right page references.
By this stage, you should have something similar to this:

Equation separated from caption by Style Separator.

The last thing that needs to be done is to align the equation number. By some quirk, it seems that, while the Style separator does allow you to combine two paragraphs into one, only the first paragraph’s tab stops seem to apply—since Microsoft has totally neglected documenting the Style separator, we are left at a loss as to whether this is by design or by oversight. And yet, if you close and save the file, and re-open it, you may find Word wanting the tab stop for the second “paragraph.” Thus, the safest thing to do seems to be to add the same tab stop on both sides of the Style separator—you need to position the I-beam both to the right and the left of the Style separator, and then add a right tab stop (remember that the Tabs dialog can easily be opened with Alt, O, T) equal to the text width of your page (e.g., if the ruler width is 16 cm, then add a 16cm right tab stop), as is demonstrated below:

Adding a tab stop to align an equation caption

The end result should be something like this:

A properly aligned equation caption

 Of course, all that work can get quite tedious, which is why I have created a simple macro to automate the work for you. You could add this yourself, or you could install my Word uTIlities, which includes this macro where it can be run with a click on the ribbon.

Sub AddCaptionedEquation()
'Created by J. Raubenheimer, 2012
'Adds an empty equation box and a right-aligned equation caption
  With Selection
    .MoveUp Unit:=wdLine, Count:=2
    .OMaths.Add .Range
    .MoveRight Unit:=wdCharacter, Count:=1
    .TypeText Text:=vbTab
    .TypeText Text:="("
    .InsertCaption Label:="Equation", ExcludeLabel:=1
    .TypeText Text:=")"
    .HomeKey Unit:=wdLine
    Dim intPos As Integer
    With .PageSetup
      intPos = .PageWidth - .LeftMargin _
      - .RightMargin - .Gutter
    End With
    .ParagraphFormat.TabStops.Add Position:=intPos, _
     Alignment:=wdAlignTabRight, Leader:=wdTabLeaderSpaces
    .HomeKey unit:=wdLine
    .ParagraphFormat.TabStops.Add Position:=intPos, _
     Alignment:=wdAlignTabRight, Leader:=wdTabLeaderSpaces
  End With
End Sub
Please note, though, that this macro must only be run in a document which already uses the Heading styles to divide and number the chapters.

One last thing about this method needs to be said. Because of the way Word handles cross-references, you will notice that a cross-reference in your text to the equation will include the parenthesis on the left of the equation number. This problem, is caused by the hidden bookmark Word adds around the caption number when you add the cross reference. Again, my uTIlities contain two tools (the Bookmark Manager or the Shrink bookmark tool) that fix this.

After all is said and done, I admit that this approach is still not ideal–the ideal would be a real fix from Microsoft–but it does, to me, present several advantages over other methods, and the fact that I have automated it means that this is not a bad solution to the student who needs numbered equations–it’s what I use.


Using BibWord for extended referencing capabilities in Word

What options does a postgraduate student who wants to automate their referencing have? Should they buy an “expensive” RMS program? Should they consider an open source alternative? Or can Word’s citation tool  still make the grade for a thesis or dissertation despite it’s shortcomings?

Enter BibWord. BibWord was developed by Yves Dhondt of MIT in an attempt to overcome some of these limitations to the Word Citations & Bibliography tool. It is downloadable freely from the website (

Essentially, BibWord allows you to “hotwire” Word’s citation styles and add styles of your own. The tool basically consists of two parts: Firstly, the “end user” tools are for students who just want to use the tools, you will find some extra styles which Yves created on the site. These can be copied to the Bibliography Styles folder on your PC  Also, there is the BibWord Extender which allows for the creation of numerically ordered in-text citations (e.g., when referencing according to the Vancouver referencing style). Secondly, there are the “developer” tools, the BibWord template and the BibType tool, for more advanced users, who wish to develop their own styles. I have, for example, used the BibWord template to develop a (relatively complete) South African Harvard style for Word.

If you want to use the developer set of tools, details can be found on the BibWord site. I will briefly show how to use the BibWord end user tools here.

The first thing to do is to copy the BibWord styles to the Style folder of your Word installation (typically, for Word 2010,
C:\Program Files\Microsoft Office\Office14\Bibliography\Style, and not
C:\Users\UserName\AppData\Roaming\Microsoft\Bibliography\Style)—you may need administrator access to do this. The figure below shows this folder on my PC (Windows 7, Office 2010, 32-bit), before the BibWord styles were added (i.e., these are the styles provided by Microsoft with Word 2010 SP1 or Word 2007 SP3).


Word Citation Styles folder

The styles that are available from the BibWord web page (as at September 2012) are listed in the table below. One more style—HarvardSA.XSL—which I created using the BibWord template, can be downloaded from the examples page.

Referencing Styles Available via BibWord

Style File names
Associação Brasileira de Normas Técnicas (ABNT) ABNT_Author.XSL
Association for Computing Machinery (ACM) ACMCitSeq.XSL
American Medical Society (ACS) ACSCitSeq.xsl
American Medical Association (AMA) AMA.XSL
American Sociological Association (ASA) ASA.XSL
Chicago Footnotes (beta – not in zip) CMSFootnote.XSL
Council of Science Editors (CSE) CSECitSeq.XSL
Harvard – AGPS HarvardAGPS.XSL
Harvard – Anglia HarvardAnglia.XSL
Harvard – Exeter HarvardExeter.XSL
Harvard – Leeds HarvardLeeds.XSL
Humana Press Humana.XSL
IEEE IEEE_Alphabetical.XSL
Lecture Notes in Computer Science (LNCS) LNCS.XSL
Modern Humanities Research Association (MHRA) MHRAFootnote.XSL
Nature Nature.XSL
Vancouver Vancouver.XSL

Once these styles have been copied to the Style folder, they will appear in the Style list of the Citations & Bibliography group of the References ribbon, and can be used as described before.


Citation Style list showing extra BibWord styles

Note, though, that if the document is opened on another computer which does not have the styles loaded, the Style will display as Unknown, and any attempt to work with the citations or bibliography will result in Word changing the style to the default style for that computer.

BibWord style not found on computer

A last lifeline from Yves is the BibWord Extender, which overcomes two problems with Word’s own referencing system. Firstly, if multiple references are used from the same author(s) in the same year, the convention is to add a date suffix (e.g., 2012a, 2012b, etc.). Word cannot do this. Secondly, while Word can number references in the order they occur in the text (when using numerical referencing styles, of course), Word cannot order the bibliography alphabetically and then assign the numbers in the text accordingly. The BibWord Extender allows both of these possibilities, and it should be noted that the BibWord styles which are suffixed with an asterisk are styles which will benefit from the use of the Extender. Unfortunately, because of the way Word creates citations, making any changes to the references (e.g., adding new sources), may invalidate what has been done by the Extender, requiring it to be rerun. Thus, it is best to use it only when the document has been finalised, and no more changes to the sources will be made. Simply start the tool (it may as for confirmation or an indication of where the bibliography styles are located on your computer), then use the file browser button to load the file to the Word Document text box, and select Extend. Once it has completed, re-open Word, and re-select the same referencing style to force an update of the citation and bibliography fields.

BibWord Extender

A nice touch of this tool, which I recommend you use, is that it first creates a backup of the Word document. The backup is indicated with an “_orig” suffix before the file extension.


So can a thesis be done with this tool? In short, yes. I remain very disappointed by a generally substandard offering from Microsoft, but if you look at the thesis.docx file on the examples page, you will see an example of a thesis that used this for its referencing, and was submitted and accepted. If you’re planning a serious, long-term academic career, then I would say: Don’t waste your time. If you have to do a thesis or even a dissertation so that you can complete your degree and then venture out into the big wide world, then this tool will do the job for you with the minimum of fuss. Just don’t try to add hundreds of references–you will find it becoming unwieldy as you try to manage your references. I suspect that one or two hundred references will be fine, but the more you add, the worse it’s lack of serious sorting, searching and grouping tools will count against it.


Struggling to view those folders mentioned above to which you need top copy the files? Then read this post: Getting proper access to all the files on your PC