How to Migrate
Your Data Out of FileCabinet CS

FileCabinet CS Migration
Outline

Every migration has six basic steps. In this guide we will cover the entire process, show you what to watch out for, and give you some tips to make it faster.

Estimating the time

FileCabinet migrations take time. Solving these two problems is where you will spend most of yours.

Problem 1

There is no bulk export

You need to convert the data from all of your drawers back into its native file type, but FileCabinet doesn't have a feature to let you do this all at once. You have to go one drawer at a time. You can't set the export to run when you leave for the day and come back the next morning to a completed job.

Problem 2

The export comes out flat

The export doesn't preserve your folder structure. Every document lands in one flat folder, and your drawer and folder organization has to be rebuilt afterward. We cover how in Step 4.

Three stats to check

1

Setup › System Configuration › Data Locations › Stats…

2

Data Location Stats: total size and drawer count per location

Estimate how long it will take

The total export time will vary based on data size and number of files. Two firms can have 100GB of client data, but one firm might have 300 drawers while another has 800. Generally the size of the data in GB is the primary driver of export time.

Runtime is not the same as your time

Example firm time estimate for Step 1

Prepping the export

Before you begin exporting your data there are a few steps to take to make the process faster, more reliable, and safe.

1

File › Backup…

2

Select All, set Focus to Active, and check “Exclude drawers unchanged since their last backup”

Extracting a drawer

FileCabinet's built-in export converts a drawer's documents back into regular files on disk. It works one drawer at a time. This is the loop you'll repeat for every drawer on your migration list.

The export loop

1. Select the client drawer you want to export.
2. Go to File › Send To › File… to open the export dialog.
3. Click Select › to add the drawer's documents to the export list on the right.
4. Check your file name options and confirm the destination folder under Modify Destination…
5. Click OK, let it run, and log the drawer as done. Repeat for the next one.

File › Send To › File… — select the documents, set the options, export

A few notes on the options

Reconstructing the folders

This is the step firms don't know exists until they get to it. FileCabinet exports everything into a flat output: one folder, thousands of files, no client subfolders. Multi-page documents may also come out as separate per-page files depending on the options you selected in the previous step (you can combine them into one, but that increases the export time). The filename options you set in Step 3 are the only thing carrying the drawer, folder, and date information out with each file.

The export folder — every document in one flat directory, one file per page, metadata packed into the filename

Three ways to rebuild the tree

For a handful of drawers, dragging files into folders is fine. At 410,000 documents, thirty seconds each is 3,400 hours. Pick the approach that matches your volume.

Automate it (Option 3)

Scripting is by far the fastest and most efficient method, but can be intimidating to tackle if you have never looked at code before. Fortunately, this is a relatively straightforward task that can be handled with a little help from modern AI tools. We have provided some example code below to help get you started along with a few things to ask.

# rebuild.py — flat export → client/folder tree
# 01234_DOE, JOHN_2025 Tax Documents_03-03-2026_030126 E-mail re Tax Info.pdf
import re, shutil
from pathlib import Path

EXPORT = Path(r"C:\Migration\Export")     # the flat export from Step 3
OUTPUT = Path(r"C:\Migration\Rebuilt")    # where the tree gets built

PATTERN = re.compile(
    r"^(?P<id>\d+)_(?P<client>.+?)_(?P<folder>[^_]+)"
    r"_(?P<date>\d{2}-\d{2}-\d{4})_(?P<doc>.+)$")

for f in EXPORT.iterdir():
    m = PATTERN.match(f.stem)
    if m:                                  # parsed: file it properly
        dest, name = OUTPUT / m["client"] / m["folder"], m["doc"] + f.suffix
    else:                                  # couldn't parse: set aside for review
        dest, name = OUTPUT / "_unparsed", f.name
    dest.mkdir(parents=True, exist_ok=True)
    shutil.copy2(f, dest / name)           # copy — the export is never touched

Verifying everything
is there

Most people skip this step, but it's the only way to know whether the previous steps worked.

1. Reconcile counts per drawer. Count the files under each rebuilt client folder and compare with the stats listed in Filecabinet. Any gap gets investigated before you go further.
2. Find the zero-byte and tiny files. Sort the whole rebuilt tree by size. Anything at 0 KB, or under about 2 KB for a PDF, did not export correctly. These are invisible in a folder listing and obvious in a size sort.
3. Spot-check with a human, across the whole date range. Open twenty documents: your oldest years, middle years, and last year. Older scans by earlier staff fail most often. Then hand a sample of rebuilt client folders to the person who knows the filing system. They will spot a structural problem in thirty seconds that no count or hash will catch.

Uploading into
the destination

Every platform wants a slightly different shape. Find out what yours wants before you upload, not after.

6.1 · Match the import shape first

Most accounting platforms build their client list and folder hierarchy from the folder structure you hand them, but they differ on the details. Does the client folder name need to match a client ID? Do year folders sit above or below document-type folders? What characters are allowed? Ask the vendor for their import spec, adjust Step 4's script to produce exactly that, and rebuild once rather than reorganize twice.

6.2 · Pilot one client end to end

Upload a single rebuilt client folder. Then, as a user: find their 2019 return, open it, search for it by name, check that the year folders are in the right order. Fix anything wrong in the script, rebuild, and only then run the full load.

6.3 · Confirm the forward workflow

FileCabinet's quiet superpower was that finished returns printed straight into the right drawer. Make sure the equivalent works on the new platform, whether that's UltraTax's print-to-DMS integration or whatever replaces it, before filing season.

6.4 · Keep the source until a full cycle has passed

Do not delete, wipe, or decommission the FileCabinet machine on the strength of a successful upload. Run both through one full season. When nobody has needed the old system for a cycle, retire it. Archive copies in two places, one offsite. After the sunset you will still be able to read your old data if you need to, so it's worth keeping around until your retention period has ended.

Where doing it yourself
makes sense

A FileCabinet migration is going to take more time than you think it will… but that may not be a problem. We have developed proprietary automated tools for every step of the process outlined in this guide. Even with those tools, migrations still take days and require manual review. In a highly seasonal business this may not be a problem. If you have the time and resources available, this is definitely a project you can handle in house. Work from a backup, figure out how to reconstruct your folders for your new software platform, and set aside some time daily to work on the migration. Over a month or two you can get it done. For everything else…

This guide is general information and does not constitute legal, technology, or professional advice for any particular firm. Product names are the property of their respective owners. Test every step against a backup before running it against your only copy.