当长文本/备注字符串是主键字段或“已索引：是（允许重复）”时，将长文本/备注字符串截断为255个字符

Question

I created a table in MS Access 2013 with only one column of "Long Text" type (called as Memo earlier) and made it the primary key of the table. I stored a long string of 255+ characters and then I tried to store another string whose first 255 characters were same as previous stored string but all other characters after first 255 were different and MS Access gave "duplicate data" error. In the new string I changed the characters that were after 255th position, using different combinations of characters and all gave error. But when I change any character before the 255th position it does not give any error. So, I concluded that MS Access checks only the first 255 characters of "Long Text" data type for checking duplicates in that column. Is it so? What else could be reason?

String Stored of 256 characters: LoremIpsumissimplydummytextoftheprintingandtypesettingindustryLoremIpsumhasbeentheindustrysstandarddummytexteversincethe1500swhenanunknownprintertookagalleyoftypeandscrambledittomakeatypespecimenbookIthassurvivednotonlyfivecenturiesbutalsotheleapintoelectr

String Gave Error: LoremIpsumissimplydummytextoftheprintingandtypesettingindustryLoremIpsumhasbeentheindustrysstandarddummytexteversincethe1500swhenanunknownprintertookagalleyoftypeandscrambledittomakeatypespecimenbookIthassurvivednotonlyfivecenturiesbutalsotheleapintoelect1

String Gave Error: LoremIpsumissimplydummytextoftheprintingandtypesettingindustryLoremIpsumhasbeentheindustrysstandarddummytexteversincethe1500swhenanunknownprintertookagalleyoftypeandscrambledittomakeatypespecimenbookIthassurvivednotonlyfivecenturiesbutalsotheleapintoelect2

String Gave Error: LoremIpsumissimplydummytextoftheprintingandtypesettingindustryLoremIpsumhasbeentheindustrysstandarddummytexteversincethe1500swhenanunknownprintertookagalleyoftypeandscrambledittomakeatypespecimenbookIthassurvivednotonlyfivecenturiesbutalsotheleapintoelect123

Does Not Give Error: LoremIpsumissimplydummytextoftheprintingandtypesettingindustryLoremIpsumhasbeentheindustrysstandarddummytexteversincethe1500swhenanunknownprintertookagalleyoftypeandscrambledittomakeatypespecimenbookIthassurvivednotonlyfivecenturiesbutalsotheleapintoelec1

Does Not Give Error: LoremIpsumissimplydummytextoftheprintingandtypesettingindustryLoremIpsumhasbeentheindustrysstandarddummytexteversincethe1500swhenanunknownprintertookagalleyoftypeandscrambledittomakeatypespecimenbookIthassurvivednotonlyfivecenturiesbutalsotheleapintoelec2

Does Not Give Error: LoremIpsumissimplydummytextoftheprintingandtypesettingindustryLoremIpsumhasbeentheindustrysstandarddummytexteversincethe1500swhenanunknownprintertookagalleyoftypeandscrambledittomakeatypespecimenbookIthassurvivednotonlyfivecenturiesbutalsotheleapintoelec3

Please notice the difference in the last few characters of above samples. The first stored string has 256 characters. Even if the column is not the primary key, the problem remains same if "Indexed: Yes (no-duplicates) allowed" value is set true in the table design for that column.

Answer 1

As @HansUp stated in the comments, Access (specifically the Jet/ACE db engine) only uses the first 255 characters of a Memo/Long Text field to create its index. Hence, it only uses the first 255 characters to enforce No Duplicates.

@HansUp's advice to use a different db engine that provides better support for long strings and Full Text search is probably the best approach, but I understand there are often other considerations that may be limiting you to solving your problem in Access.

As such, here is an Access-only approach to solving your problem. This assumes the requirement you listed in the comments is valid; ie, you need to store unique strings of between 400 and 1000 characters.

---

Alternative 1

1. Keep your initial Memo/Long Text field: Notes
2. Create four text fields (not Memo/Long Text) of 250 characters max: Notes1, Notes2, Notes3, Notes4
3. Set all four text fields: Required -> True and Allow Zero Length -> True (this is required to ensure the unique index is enforced for strings less than 751 characters)
4. Create a unique index and add all four text fields to that index
5. Don't ignore nulls in your index
6. When you store the values, you will need to store them in the Notes field and also split the string among the four smaller NotesX fields

---

Alternative 2:

Keep your current setup and enforce the uniqueness at code level. Every time you update or insert a note, do a search on all notes that match the first 255 characters, read the value and perform the comparison in code.

---

Alternative 3 ( thanks to @HansUp for suggesting this in the comments ):

1. Keep your initial Memo/Long Text field: Notes
2. Create a 16 or 32 character text field to store the 256 bit or 512 bit hash of your long text: NotesHash
3. Add a unique index to your NotesHash field
4. Every time the memo field is changed, re-compute the hash value and attempt to store it in the table

Notes for this method :

• As the pigeonhole principle easily proves, there is the possibility that two different strings will generate the same hash (a collision). However, using a good hashing algorithm will make the actual probability approach zero.
• This site offers some VB6/VBA/VBScript implementations of various hashing algorithms. I can't vouch for their correctness, but they passed the eye test for me. Use at your own risk, but it's at least a good starting point.
• Really, you can use any deterministic function that returns a string of 255 characters or fewer given an arbitrarily large input. The difference between a crappy hash algorithm and a good one is how well it minimizes collisions. For that reason, I would suggest you use one based on a popular standard.

---

And yes, I still highly recommend @HansUp's solution to simply use a different db engine.