OmniDocWord extracts specified information into Excel in batches, first determine the field table
To organize a batch of Word documents with similar formats into an Excel table, the key is to determine the field table before starting. This article gives how to write the field table, sample confirmation method, extraction steps and item-by-item checklist, without relying on a special tool.
Key points of this article
- Write a field table before you start, and clarify the field name, location and format. Subsequent extraction and verification will be based on it.
- First select a sample to confirm the field positions, and then apply the same rule to all documents. Documents with inconsistent structures are processed separately.
- The Excel header corresponds to the field table one-to-one, and a "source file name" column is added to facilitate checking the original text after making an error.
- Randomly check more than five copies and compare them box by box with the original text, focusing on the amount, date and number.
- If a field is missing from the original text, mark it as"not filled in". Do not leave it blank or make up the number to avoid misreading.
If what you want to do is to extract some information in Word into Excel in batches, such as extracting the contracting parties, amounts and signing dates from dozens of contracts and organizing them into a table, the problem is often not the copying and pasting itself, but the fact that the fields are not clearly defined until halfway through: some documents are missing certain items, some numbers are misplaced, and you don’t know which ones have been processed when you rework.
A more stable approach is to first write a field table, determine the name, position and format of each field in the document, then extract them uniformly based on this table, and finally use the Excel header to take over the results. This idea does not rely on a specific batch tool. You can do it if you have Word and Excel on hand. OmniDoc is responsible for opening, viewing and organizing document content, and helping you put fields into tables.
How to extract information in batches from a batch of Word documents? One sentence conclusion
Define the field table first, then execute it using the same set of rules, and finally check it item by item. Tools are just a means to execute the field table. If the field table is not clearly defined, the results of any tool will not be reliable.
| your situation | Recommended practices | reason |
|---|---|---|
| There are not many documents (less than ten), and the structure of each document is quite different. | Open one by one and manually fill in Excel according to the field table | When the structure is not unified, rule extraction is prone to errors, but manual verification is faster. |
| Dozens or more copies are generated from the same template | First determine the location in a sample, then apply the same rule to all documents | When the templates are unified, the field positions must be consistent so that rules can be reused. |
| Information is scattered in the text, tables, headers and footers | Fields are recorded separately according to location category, and the table header is also designed according to category. | The content in different locations is taken in different ways, and it is easy to miss when mixed together. |
| The source file is a scanned PDF, not Word | Do text recognition first, confirm that the text can be selected, and then follow the process of this article. | The content of text that cannot be selected cannot be directly searched for in the field. |
| Need to be repeated monthly or quarterly | Fix the field table and header as a template, and only replace the source file each time | After the template is stabilized, the scope of each verification is fixed. |
Before you start: Confirm these points first
- Is the field really needed: List the information that readers really need, such as contract number, contracting party, amount, and signing date. The fewer fields, the faster the verification. Keep the ones you are sure to use first, and don’t be greedy for the rest.
- Whether the sample documents are uniform: Open three to five copies at random to see if the contents such as "Party A" and "Amount" are in similar positions and written in similar ways. If they are not unified, separate rules must be written for the variants, otherwise the entire batch of data will be quietly misaligned.
- Can text be selected: Word documents can usually find and select text directly. If the source is a PDF, try to select a piece of text first. If you cannot select it, it means that the content is an image and needs to be recognized first.
- How to design the header: The header name corresponds to the field table one-to-one, the date is in a unified format, and the amount is unified as a number. If the table header is set early, the subsequent data will not be misaligned.
- Whether to keep the original: The source file may be accidentally modified during extraction. Make a working copy first, and all operations will be performed on the copy.
Steps to organize Word information by field table
- Write out the field table: Create a new worksheet in Excel. Write the field name in column A, write the location in column B (such as "first table on the home page" and "signing place"), and write the format requirements in column C (such as "YYYY-MM-DD").
- Pick a sample to confirm the position: Open a representative document, find the content field by field, and fill in the actual position and writing method into the field table. If you encounter exceptions, write them down in the remarks column.
- Open the document in OmniDoc for verification: Open the Word document in app.unidoc.top, directly view the text, table and header content, and confirm that the sample matches the field table.
- Create the header in Excel: The order of the header is consistent with the field table. Each column corresponds to a field. First fill in a row of sample data for testing.
- Extract one by one according to the same rules: For each document, take the value from the field table and fill in a new row. When there are a large number of files, you can ask colleagues who know scripts to use a common method to read the docx files and execute them uniformly, provided that the sample rules have been verified.
- Save in batches and record sources: Save each batch after processing, and add a column of "source file name" to facilitate review.
- Uniform format: Date, amount, and number are organized into the same format, removing extra spaces and line breaks.
- Save as final table: Save as an Excel workbook. When you need to send it to others for review, export another PDF version, and the original table will still remain in editable format.
Check each item after completion
| Check items | How to see | What to do if there is a problem |
|---|---|---|
| Is the number of lines consistent with the number of files? | Compare the number of files in a folder with the number of table rows | Find missing files, fill them in one by one and record the reasons |
| Whether the field is misplaced | More than five samples were randomly selected and compared with the original text one by one | Modify the field location rules to re-extract similar files |
| Number and date format | Check the decimal point of the amount and the order of the date. | Unified correction with find replace or break down function |
| Blank and abnormal value | Filter out the blank space to confirm whether the original text does not have this information. | If the original text is missing, it will be marked "not filled in". Do not leave it blank or add up the number. |
| Can the source be traced back? | Whether the source file name of each line is accurate | Make up the source column to avoid not being able to check the original text afterwards |
# frequently asked questions (FAQ)
1. The information in the Word document is in the table. Can I copy the whole piece to Excel?
Yes, but merging cells can easily lead to dislocation. Cancel the unnecessary merge in Word before copying, or copy only a few columns you need, paste them into Excel and then sort them out.
2. There is a number and date in the file name. can you extract it together?
Sure. Record the number and date in the file name as independent fields, and the "location" column of the field table is marked as "file name", which is checked separately with the content of the text to avoid two places of information covering each other.
3. What if the same field is written differently in different documents?
Include all common words in the field table notes, for example, "amount" may be written as "total" or "total price". Determine the correspondence manually, and don't let the differences in writing sneak into the table.
4. Can the extraction results be directly used for formal submission?
It is recommended to conduct a spot check first and confirm the key information such as the amount and the signatory against the original text before using it. The content of OCR or machine-aided identification especially needs to be checked manually.
5. How to extract the information from the scanned piece?
First judge whether the text can be selected. When the selection is not correct, it is necessary to do character recognition first, and then extract it according to this process, and the recognition results should also be compared with the original item by item.
# Summary
First set the field list, then execute it uniformly, and finally check it item by item. Write down the fields, locations, and formats clearly, and the batch extraction of some information by word to excel will become a process that can be reused, rather than manual work that can be redone over and over again.
