How do I convert from BLOB to TEXT in MySQL?

Question

I have a whole lot of records where text has been stored in a blob in MySQL. For ease of handling I'd like to change the format in the database to TEXT... Any ideas how easily to make the change so as not to interrupt the data - I guess it will need to be encoded properly?

Dismiss

Answer 1

That's unnecessary. Just use SELECT CONVERT(column USING utf8) FROM..... instead of just SELECT column FROM...

Retread

Usage: SELECT CONVERT(column USING utf8) FROM table; – Institutionalize

This works great for those GROUP_CONCATs that convert your output to blobs and you really want them as strings. I had an issue similar to the OP's while using Node.JS with the node-mysql library - this fixed all group_concat issues. – Fertilizer

This work, and also can use with hot query like CONVERT(LEFT(MD5([ID]),8) USING utf8) – Thitherto

This won't work. The charset needs to be utf16 or else it will result in data loss if it encounters a set of bytes that can't be converted to utf8. It will replace those bytes with a ? character resulting in data loss. – Dumpish

CONVERT(column USING utf8) will return a column with the name CONVERT(...). If one would like to avoid this, don't forget using CONVERT(column USING utf8) AS column. This will rename the column to column. – Hultin

Answer 2

Here's an example of a person who wants to convert a blob to char(1000) with UTF-8 encoding:

CAST(a.ar_options AS CHAR(10000) CHARACTER SET utf8)

This is his answer. There is probably much more you can read about CAST right here. I hope it helps some.

Celaeno

Unfortunately this isnt working for me. I get empty rows and sometimes only a 1-character-output with strange symbols. – Hedvig

Worked in select query as well, select A.id,CAST(B.content AS CHAR(10000) CHARACTER SET utf8) as content, B.b from A join B ON B.content_id=A.content_id – Jezabelle

Worked for me while trying to JSON_EXTRACT on a BLOB type column :) SELECT ... CAST(JSON_EXTRACT(column, '$.field') AS CHAR(10000) CHARACTER SET utf8) FROM ... – Nyssa

Answer 3

I have had the same problem, and here is my solution:

1. create new columns of type text in the table for each blob column
2. convert all the blobs to text and save them in the new columns
3. remove the blob columns
4. rename the new columns to the names of the removed ones

ALTER TABLE mytable
ADD COLUMN field1_new TEXT NOT NULL,
ADD COLUMN field2_new TEXT NOT NULL;

update mytable set
field1_new = CONVERT(field1 USING utf8),
field2_new = CONVERT(field2 USING utf8);

alter table mytable
drop column field1,
drop column field2;

alter table mytable
change column field1_new field1 text,
change column field2_new field2 text;

Weasand

This is the only answer that worked for me, thanks:) – Lello

The intermediate column did the trick. Kept getting errors citing bad characters via all other methods and answers. Thanks – Racecourse

Answer 4

If you are using MYSQL-WORKBENCH, then you can select blob column normally and right click on column and click open value in editor. refer screenshot:

Tenfold

FYI: (So that people don't have to click on the picture above to see what this does): this shows ONE value at a time - doesn't solve the desire to see text for that column in all rows. – Sisyphus

Answer 5

None of these answers worked for me. When converting to UTF8, when the encoder encounters a set of bytes it can't convert to UTF8 it will result in a ? substitution which results in data loss. You need to use UTF16:

SELECT
    blobfield,
    CONVERT(blobfield USING utf16),
    CONVERT(CONVERT(blobfield USING utf16), BINARY),
    CAST(blobfield  AS CHAR(10000) CHARACTER SET utf16),
    CAST(CAST(blobfield  AS CHAR(10000) CHARACTER SET utf16) AS BINARY)

You can inspect the binary values in MySQL Workbench. Right click on the field -> Open Value in Viewer-> Binary. When converted back to BINARY the binary values should be the same as the original.

Alternatively, you can just use base-64 which was made for this purpose:

SELECT
    blobfield,
    TO_BASE64(blobfield),
    FROM_BASE64(TO_BASE64(blobfield))

Dumpish

NOTE: TO_BASE64 has a different purpose than the original post's question (OP). OP "knows" they have text stored into a blob field, and wants to see that text as text. (For that purpose, the conversion to utf16 is appropriate - if the text was originally in utf16). TO_BASE64 is a way to show an arbitrary blob using visible characters. It expands the original bytes, using a set of 64 characters. – Sisyphus

MORE: If the data was originally in utf8 (from whatever source it was received, then written into the blob), utf16 will not do what you want - you need to "know" what format the original text was in, and use that in the USING clause. For example, my general log table stores queries that were utf8, USING utf16 on argument column, resulted in a sequence of Asian characters, for simple queries like SHOW SESSION VARIABLES. In my case USING utf8 was the correct answer. – Sisyphus

Answer 6

You can do it very easily.

ALTER TABLE `table_name` CHANGE COLUMN `column_name` `column_name` LONGTEXT NULL DEFAULT NULL ;

The above query worked for me. I hope it helps you too.

Oceania

Answer 7

Using phpMyAdmin you can also set the options to show BLOB content and show complete text.

Unseen

How exactly would someone go about doing this? Your answer will be much more helpful if you post some code and/or screenshots detailing such. – Rhinestone

Answer 8

Or you can use this function:

DELIMITER $$

CREATE FUNCTION BLOB2TXT (blobfield VARCHAR(255)) RETURNS longtext
DETERMINISTIC
NO SQL
BEGIN
       RETURN CAST(blobfield AS CHAR(10000) CHARACTER SET utf8);
END
$$


DELIMITER ;

Zrike

Answer 9

I had the same issue with my MariaDB records. It was solved (by my colleague) using

select
uncompress(blobfield)
from table

Eudoca

Answer 10

SELECCT TO_BASE64(blobfield)  
FROM the Table

worked for me.

The CAST(blobfield AS CHAR(10000) CHARACTER SET utf8) and CAST(blobfield AS CHAR(10000) CHARACTER SET utf16) did not show me the text value I wanted to get.

Broadside

Answer 11

i cannot understand why not so easy:

UUID_TO_BIN('77dea2ad-3c8c-40c6-a278-7cf1a1ac9384')

and

BIN_TO_UUID(FIELD)

Slipcase