CDLI Search Information

The CDLI online database contains, as of 15 November 2010, catalogue information for approximately 255,000 cuneiform texts, 118,000 images, and 1.5 million lines of transliteration. Search functions of the website are powered by a MySQL database developed by UCLA computer science graduate students Marjan Yahyanejad, Saurabh Trivedi, and Oren Freiberg.

Searching the Catalogue

When searching for tablets in the catalogue, it is possible to choose between three modes of display as well as the number of tablets to be displayed.

  • Full search displays, in the first of three columns, an essentials catalogue, in the second a preferred thunbnail (expandable) of the cuneiform artifact together with hyperlinks to further available images, and in the third a text transliteration, in some cases with interlinear translations. Where user search includes transliteration or translation, those items are color coded to simplify their location in the full text. It is possible to cross-search the dataset of transliterations for words and graphemes, limited by period, provenience, etc., as well as to sort results according to catalogue entries.
  • Concise search displays a simple table of found texts with essential catalogue information, in the case links to the CDLI pages with images and transliteration. Where user search includes transliteration or translation, concise returns only the line of text containing those items, that are also color coded. Table fields in the results pages are sortable.
  • Browse search displays full text in single page with extended catalogue, all available image thumbnails and transliterations. Navigate with “left and right” arrows in page headers to view preceding, and following entries.

CDLI search page defaults to settings for full search and for maximally 1000 hits to be displayed per web page.

Search Tips

  • Many museum and publication numbers contain leading zeros in the database so that records sort properly. For example, TRU 1 is in the database as TRU 001. If something you expect to find does not show up, try a short version (here, “TRU”) to see how the entries are numbered.

Simple Search of Words or Parts of Words in Transliterations

Search in transliterations can be set to locate one or more signs or words in an exact string, or across a full line or text:

  • Select Full, concise or browse from the represtentation menu, and type your search item(s) into the transliteration box, following CDLI ATF-transliteration conventions (š is coded “sz”, emphatics ṣ and áš­ are coded “s,” and “t,”, and ḫ is simple “h”; for further conventions, see the ATF conventions pages used by CDLI and Oracc).
  • Clicking “Exact search” returns only those instances where the uninterrupted string is located in transliterations; ATF editing characters, for instance “#” to indicate a damaged sign, or “!” to indicate an emendation, are ignored in default setting of CDLI search. User can restrict search to those instances including such characters by clicking on the corresponding button below the transliteration search box.
  • Optional: Restrict the search by typing additional conditions into the catalogue fields.
  • Hit your keyboard’s return, or click on the search button at the end of the page to initiate your search.

You will receive a list of occurrences that offers you some options for further discovery:

  • In concise search, click on the column headers to sort occording to those catalogue criteria.
  • You may further expand your results to show full transliterations and avialable images and image links by clicking on the corresponding hyperlink above your search results. In the case of multiple item transliteration search, you may also expand the results to display all instances of full text transliterations that included searched items somewhere in the text, not just in one line.
  • If more occurences have been found than are displayed, click on the next page link above or below the results to see the remaining items.

Some Useful Explanations of the Simple Search Procedure

When searching for words or parts of words in the transliterations, you should be acquainted with the ATF conventions that apply to all texts in the CDLI and Oracc corpora.

It is useful to understand how the simple search procedure works.

Searching for words:

  • A word has to be written as a set of one or more graphemes separated by dashes. Determinatives (semantic or phonetic glosses such as “d” = “divinity”, “ki” = “place”) are to be enclosed in curly brackets { and }. The special character “š” in upper or lower case (Š, š) is coded in CDLI ASCII with “SZ” and “sz”, respectively. There is no special character for “ḫ” that is either “H” or “h”.
  • CDLI search .
  • Subsequently, searches are performed on the transliterations of these tablets in order to identify the lines containing the word searched for.
  • Finally, the designations of the found tablets are displayed together with transliterations of the lines containing this word.

Searching for parts of words:

  • Searching for an individual grapheme will result in precisely those lines containing this grapheme. For instance, searching for lu2 (human) as part of a word will find all tablets with the word lu2, but also, e.g., lu2-kal-la (an official at Umma) or lu2-u18 (“mankind”).
  • Searching for part of a word consisting of a sequence of graphemes will result in a list of tablets that contain this sequence.

Search Tips

  • Servers invariably experience periods of high traffic, or a software break-dowon that causes a website to slow down or break apart. We regularly monitor CDLI for such interruptions, but are grateful for user reports of problems.
  • We are working to normalize all sign readings in CDLI, but we deal with quite a lot of legacy transliterations with varying standards. You therefore might want to search independently for “gisz”, “gesz”, and “GISZ” to find all ocurrences of the grapheme. Similarly, since the archaic texts (ca. 3400-2700 BC) are coded according to standards of the Uruk Project and the sign list ATU 2, you have to search independently for “dug”, “DUG~a”, “DUG~b”, etc.