Epicor DMT (Data Management Tool): Bulk Import Constraints, Limits, and Best Practices

TL;DR

• Bottom line: Epicor DMT processes records through business objects one-at-a-time (not direct SQL), so bulk imports are inherently slow -- expect 30-90 records per minute depending on table complexity and BPMs. Split large files into 10,000-record chunks and run parallel DMT instances to dramatically reduce total import time.
• Key limit: No hard record-count cap per file, but practical throughput is 30-90 RPM. A 100,000-record import takes 18-55 hours single-threaded. With 10 parallel instances, this drops to under 5 hours.
• Watch out for: BPMs (Business Process Management rules) fire on every DMT record. Badly designed BPMs can double your import time. Disable or condition them before bulk loads.
• Best for: One-time data migrations (go-live cutover), bulk master data loads (Parts, Customers, Vendors, BOMs), and recurring batch updates via CLI/playlist automation.
• Authentication: DMT uses Epicor user credentials (username/password) passed via CLI arguments or entered in the GUI. No OAuth or API key support.

System Profile

This card covers the Epicor Data Management Tool (DMT), a desktop client application bundled with Epicor ERP (version 10.2.500+) and Epicor Kinetic. DMT is the official tool for bulk importing, updating, and deleting records in Epicor ERP using CSV files. It processes data through Epicor's business objects (BOs), meaning all standard validation rules, BPMs, and data directives execute on every imported record. DMT is NOT a direct database import tool -- it does not bypass business logic. [src1, src2]

API Surfaces & Capabilities

DMT is not a traditional API -- it is a client-side tool that internally calls Epicor's business objects. It exposes multiple interaction surfaces for automation. [src1]

Rate Limits & Quotas

Per-Record Processing Limits

File & Data Limits

Practical Throughput Estimates

Authentication

Authentication Gotchas

• CLI credentials are in plaintext: the -Pass= argument is visible in Task Manager and batch file contents. Restrict file permissions on automation scripts. [src1]
• DMT user needs appropriate Epicor security roles for the specific business objects being imported. [src5]
• Cloud deployments may require VPN or specific network access to reach the Epicor application server. [src2]

Constraints

• Records processed through business objects, not direct SQL -- every record triggers full field validation, required field checks, cross-reference lookups, and all active BPMs. Throughput: 30-90 RPM. [src3]
• BPMs fire on every record -- email notifications, workflow triggers, and custom logic all execute during DMT imports. A badly-designed BPM can cut throughput in half. [src3]
• Dependency order is mandatory -- you cannot import child records before parent records (e.g., Parts before BOMs, Customers before Sales Orders). [src5]
• Server Processing is limited to select tables -- the 4x-faster server-side batch processing (2024.1+) only works on a few tables. Most common import tables do not support it yet. [src4]
• Cloud tenancy restrictions -- cloud-hosted instances cannot use certain server-side optimizations available to on-premise. [src3]
• No built-in delta/incremental import -- DMT does not natively compare source data to existing Epicor data. Manage add vs. update logic yourself. [src5]
• CSV only for automation -- PlayLists and CLI require CSV format. Excel files must be saved as CSV first. [src1]
• 60+ templates available -- but not all Epicor tables have DMT templates. Custom UD tables require manual configuration. [src2]

Integration Pattern Decision Tree

START - Bulk import data into Epicor via DMT
|-- How many records?
|   |-- < 1,000 records
|   |   --> DMT GUI: interactive mode, preview validation, single file
|   |-- 1,000 - 10,000 records
|   |   --> DMT CLI: single instance, -NoUI for unattended
|   |-- 10,000 - 100,000 records
|   |   |-- Split into 10,000-record CSV files
|   |   |-- Launch 5-10 parallel DMT.exe instances
|   |   --> Disable non-essential BPMs first
|   |-- > 100,000 records
|       |-- Split into 10,000-record chunks
|       |-- Run across multiple client machines
|       --> Plan for 4-11 hours with 10 parallel instances
|-- What operation?
|   |-- Add new --> Check "Add", ensure parent records exist
|   |-- Update existing --> Check "Update", match on primary keys
|   |-- Delete --> Check "Delete", backup first
|   |-- Mixed --> Check both "Add" and "Update"
|-- One-time or recurring?
|   |-- One-time (go-live) --> Playbooks, pre-load + delta at cutover
|   |-- Recurring --> CLI + Windows Task Scheduler + error monitoring
|-- Load order (mandatory dependency sequence):
    1. Company, Sites, Plants
    2. Chart of Accounts, GL Accounts
    3. Customers, Vendors, Suppliers
    4. Part Classes, Part Groups
    5. Parts (Part Master)
    6. Part Revisions
    7. Bill of Operations (BOO)
    8. Bill of Materials (BOM)
    9. Inventory (initial balances)
    10. Open Orders, POs, Jobs

Quick Reference

Step-by-Step Integration Guide

1. Prepare CSV using DMT Template Builder

Open DMT, select the target table, click "Template Builder." Required fields are pre-checked. Export the empty CSV template and populate with your data. [src5]

# Launch DMT GUI to build template
start C:\Epicor\ERP10\DMT.exe
# In DMT GUI: select table > Template Builder > check fields > Save CSV

Verify: Click "Validate Columns" in DMT to confirm all required fields are present.

2. Clean and format data

Ensure all data meets DMT formatting requirements: MM/DD/YYYY dates, no thousand separators, UTF-8 encoding, trimmed whitespace. [src6]

import pandas as pd

df = pd.read_csv("raw_data.csv", encoding="utf-8")
# Fix dates to MM/DD/YYYY
for col in df.select_dtypes(include=["datetime64"]).columns:
    df[col] = pd.to_datetime(df[col]).dt.strftime("%m/%d/%Y")
# Remove thousand separators, trim whitespace
for col in df.select_dtypes(include=["object"]).columns:
    df[col] = df[col].str.replace(",", "", regex=False).str.strip()
df.to_csv("cleaned_for_dmt.csv", index=False, encoding="utf-8")

Verify: Dates are MM/DD/YYYY, no commas in numbers, no leading/trailing whitespace.

3. Split large files for parallel processing

For datasets over 10,000 records, split into chunks for parallel DMT execution. [src3]

import pandas as pd, math
df = pd.read_csv("cleaned_for_dmt.csv", encoding="utf-8")
chunk_size = 10000
for i in range(math.ceil(len(df) / chunk_size)):
    chunk = df[i * chunk_size : (i + 1) * chunk_size]
    chunk.to_csv(f"dmt_chunk_{i+1:03d}.csv", index=False, encoding="utf-8")

Verify: Each chunk has correct headers. Total rows across chunks = original file.

4. Disable non-essential BPMs

Identify and disable BPMs that slow imports (email notifications, workflow triggers). [src3]

In Epicor Kinetic:
1. System Setup > Business Process Management
2. Identify BPMs on target table
3. Disable non-essential BPMs or add condition: "Client Type <> DMT"
4. Document which BPMs were disabled

Verify: Test 10-record import. Target speed: 55-90 RPM without BPMs.

5. Execute parallel imports via CLI

Launch multiple DMT instances simultaneously, each processing a different chunk. [src3]

@echo off
start "" "C:\Epicor\ERP10\DMT.exe" -User=dmtuser -Pass=pass -Source="C:\data\dmt_chunk_001.csv" -Add -NoUI
start "" "C:\Epicor\ERP10\DMT.exe" -User=dmtuser -Pass=pass -Source="C:\data\dmt_chunk_002.csv" -Add -NoUI
REM ... repeat for all chunks
echo All DMT instances launched. Monitor Task Manager for completion.

Verify: Monitor Task Manager for DMT.exe processes. Check log files for errors.

6. Validate and re-enable BPMs

After imports complete, verify data integrity and restore BPMs. [src5]

1. Run BAQ queries to count imported records
2. Compare counts against source data
3. Spot-check 10-20 records for field accuracy
4. Re-enable all disabled BPMs
5. Review DMT error logs, re-import failed records

Verify: Record counts match source. Error log shows zero or expected failures. BPMs re-enabled.

Code Examples

Python: Automated parallel DMT execution

# Input:  Directory of chunk CSV files, DMT.exe path, credentials
# Output: Parallel DMT execution with completion monitoring

import subprocess, os, time, glob

DMT_PATH = r"C:\Epicor\ERP10\DMT.exe"
DATA_DIR = r"C:\data\dmt_chunks"
MAX_PARALLEL = 10

chunk_files = sorted(glob.glob(os.path.join(DATA_DIR, "dmt_chunk_*.csv")))
processes = []

for csv_file in chunk_files:
    while len([p for p in processes if p.poll() is None]) >= MAX_PARALLEL:
        time.sleep(5)
    cmd = [DMT_PATH, "-User=dmtuser", "-Pass=pass", f"-Source={csv_file}", "-Add", "-NoUI"]
    proc = subprocess.Popen(cmd)
    processes.append(proc)
    print(f"Launched: {os.path.basename(csv_file)} (PID: {proc.pid})")

for proc in processes:
    proc.wait()
print("All DMT imports complete.")

PowerShell: PlayList-based migration

# Input:  PlayList CSV with import definitions
# Output: Sequential DMT imports

$dmtPath = "C:\Epicor\ERP10\DMT.exe"
$playlistPath = "C:\data\playlist.csv"

$startTime = Get-Date
$proc = Start-Process -FilePath $dmtPath `
    -ArgumentList "-User=dmtuser", "-Pass=pass", "-PlayList=$playlistPath", "-NoUI" `
    -Wait -PassThru

$duration = (Get-Date) - $startTime
Write-Host "Completed in $($duration.TotalMinutes) minutes. Exit: $($proc.ExitCode)"

Batch: BAQ Export for backup

@echo off
REM Export current data before bulk operations
start /wait "" "C:\Epicor\ERP10\DMT.exe" -User=admin -Pass=pass ^
    -Export -BAQ="PartExport" -Target="C:\backup\parts.csv" -UseFieldNames -NoUI
echo Export complete.

Data Mapping

DMT CSV Column Requirements

Data Type Gotchas

• Lookup values must exactly match existing Epicor records -- mismatched case causes "Field references invalid value" errors. [src6]
• Empty numeric fields should use 0, not blank -- blank values cause type conversion errors. [src6]
• UOM codes must exist in Epicor before import -- import UOMs first. [src5]
• Site/Plant codes are required on most transactional records, even for single-site installations. [src5]

Error Handling & Failure Points

Common Error Messages

Failure Points in Production

• BPMs sending emails on every record: Generates thousands of emails, degrades mail server and DMT performance. Fix: Add BPM condition: skip when Client Type = DMT, or disable email BPMs during bulk loads. [src3]
• Dependency order violation: Importing BOMs before Parts causes every record to fail with "Field references invalid value." Fix: Always load in dependency order. Use Playbooks to enforce sequence. [src5]
• DMT.exe hangs after completion: When launched from .NET with RedirectStandardOutput=true, a deadlock occurs on large datasets. Fix: Set RedirectStandardOutput = false or asynchronously read stdout. [src3]
• CSV encoding issues: Non-UTF-8 files cause garbled text. Fix: Save as CSV UTF-8 in Excel. Verify encoding before import. [src6]
• Client runs out of memory: Too many parallel DMT instances crash the client. Fix: Limit to 8-10 instances per machine. Distribute across multiple machines. [src3]

Anti-Patterns

Wrong: Loading all 100K records in a single instance

REM BAD -- single instance for 100K records takes 18-55 hours
DMT.exe -User=admin -Pass=pass -Source="all_100k_parts.csv" -Add -NoUI

Correct: Split into chunks and run parallel

REM GOOD -- 10 parallel instances, each 10K records. Total: 2-5 hours.
for /L %%i in (1,1,10) do (
    start "" DMT.exe -User=admin -Pass=pass -Source="chunk_%%i.csv" -Add -NoUI
)

Wrong: Leaving all BPMs active during bulk import

REM BAD -- email BPM fires 100,000 times. Speed drops to 30 RPM.
REM Mail server queue backs up for hours.

Correct: Disable or condition BPMs first

REM GOOD -- add BPM condition: skip when CallContext.ClientType == "DMT"
REM Speed stays at 55-90 RPM. Re-enable after import.

Wrong: Importing child records before parent records

REM BAD -- BOMs before Parts. Every record fails.
DMT.exe -Source="bom_materials.csv" -Add -NoUI
DMT.exe -Source="parts.csv" -Add -NoUI

Correct: Use Playbook with dependency order

REM GOOD -- Playbook enforces: Parts > Revisions > Operations > BOMs
DMT.exe -User=admin -Pass=pass -PlayBook="migration_playbook.txt" -NoUI

Common Pitfalls

• Not testing with a small batch first: A formatting error in one column wastes hours of processing. Fix: Always test with 50-100 records first before full import. [src5]
• Using Excel format instead of CSV: DMT requires .csv for automation. Fix: Save As > CSV (Comma delimited) or CSV UTF-8 in Excel. [src1]
• Skipping column validation: Missing or misspelled columns cause batch failures. Fix: Click Validate Columns hyperlink after loading CSV. [src5]
• No backup before bulk delete: There is no undo for DMT deletes. Fix: Export via BAQ Export before any delete operation. [src1]
• Running DMT during business hours: Bulk imports consume server resources and slow interactive users. Fix: Schedule during off-hours. Coordinate with Epicor admins. [src7]
• Version mismatch between DMT and server: Field mapping and validation failures. Fix: Use DMT version matching your Epicor build. Included since 10.2.500. [src1]

Diagnostic Commands

REM Check DMT version
"C:\Epicor\ERP10\DMT.exe" -Version

REM Test connectivity with single record
"C:\Epicor\ERP10\DMT.exe" -User=test -Pass=pass -Source="single_record.csv" -Add -NoUI

REM Export data for verification
"C:\Epicor\ERP10\DMT.exe" -User=admin -Pass=pass -Export -BAQ="PartCount" -Target="verify.csv" -UseFieldNames -NoUI

REM Check DMT log files
dir "C:\Users\%USERNAME%\AppData\Local\Epicor\DMT\Logs\"

REM Monitor running DMT processes
tasklist /fi "imagename eq DMT.exe"

Version History & Compatibility

When to Use / When Not to Use

Important Caveats

• DMT throughput (30-90 RPM) is from community benchmarks, not official Epicor specifications. Actual rates vary by server hardware, network latency, table complexity, and BPM count. Always benchmark in your environment.
• Server Processing (2024.1+) is being rolled out incrementally. As of early 2026, most common high-volume tables do not support it. Check your version's release notes.
• Cloud-hosted instances may have additional throughput constraints from the hosting infrastructure not documented as DMT-specific limits.
• DMT is a Windows desktop application only. No Linux, macOS, or web-based version exists. Automation requires a Windows machine with DMT installed.
• The 60+ template count is approximate and varies by Epicor version and installed modules.