Ibrahim 

You need it create it yourself.  Auto-Suggest Dictionaries are generated from your Translation Memory, they are not something readily available off the shelf in all the language pair combinations that are possible.

Hi Paul,

How can I create it?

Ibrahim 

Go to your Translation Memories View in Trados Studio, click on "Create AutoSuggest Dictionary", select your Translation Memory in the wizard and follow the stops.

Note that you will need a translation memory with a minimum of 10,000 translation units.  This is the threshold at which the software allows the generation of AutoSuggest dictionaries to be productive.  If your TM has fewer than 10,000 TUs, the option to create an AutoSuggest dictionary may be disabled.

Thanks Paul for your help

I tryed it. The TM is under 10000 TUs. Where can I get TM larger than 10000 TUs?

 Ibrahim 

TM's are normally personalised with the material you translate.  So you would need to translate more!

But if you desperately just want to build a larger TM and don't care about the quality or context being something you have personally checked yourself... in which case machine translation might be a better option anyway... then try getting a corpus from here: https://opus.nlpl.eu/

Minimising the size... because you won't want the excessive sizes available here... you still have some massive corpus that you could take and then cut them down if you have a decent text editor so something that will get you the sufficient number of TUs you need.  I think if I was you and till thought this was an appropriate thing to do then I'd download one of these TMX files, cut it down so I only had a fraction of the number available, and then use that to get a start with AutoSuggest.  Your biggest problem, based on your comments so far, is likely to be just how to handle such large files in any text editor!!

But be very sure this is really what you need!

How large are your TMs so far?  Perhaps a better solution would be to simply machine translate some content that is relevant to the sort of work you do, post-edit it, and use that to get to 10000 translation units.  I imagine you could get there quite quickly.

 Ibrahim 

TM's are normally personalised with the material you translate.  So you would need to translate more!

But if you desperately just want to build a larger TM and don't care about the quality or context being something you have personally checked yourself... in which case machine translation might be a better option anyway... then try getting a corpus from here: https://opus.nlpl.eu/

Minimising the size... because you won't want the excessive sizes available here... you still have some massive corpus that you could take and then cut them down if you have a decent text editor so something that will get you the sufficient number of TUs you need.  I think if I was you and till thought this was an appropriate thing to do then I'd download one of these TMX files, cut it down so I only had a fraction of the number available, and then use that to get a start with AutoSuggest.  Your biggest problem, based on your comments so far, is likely to be just how to handle such large files in any text editor!!

But be very sure this is really what you need!

How large are your TMs so far?  Perhaps a better solution would be to simply machine translate some content that is relevant to the sort of work you do, post-edit it, and use that to get to 10000 translation units.  I imagine you could get there quite quickly.