The ePSD search is intended to be a DWIM (Do What
I Mean) interface--in general you can type any kind of ePSD key term
in this box and see what happens. This means citation forms,
transliteration, English words, Akkadian words. You can also enter a
fuzzy CFGW pair here in which the GW part does not need to be a
literal GW, but can be any key word from the actual GW or meanings
entries of the article, e.g., sag[good]
.
Two wildcards are available for fuzzier searches.
The asterisk wildcard, *
, matches any string and can be
used to select multiple keys, e.g., ban*
selects
ban, ban2, ... banda...
etc. It can be used anywhere in
the term, and may be used more than once.
A special-purpose wildcard is provided by the percent sign,
%
. This matches any numbers at the end of a key; use
this when you want to match all graphemes with the same basename,
e.g., kici%
selects kici13, kici14
etc.
Aliasing or fuzzy key matching is
enabled by default for graphemic searches. In this mode, a search
term selects not only the grapheme itself, but also those which the
aliases table (which you can read
here) considers to be equivalent; in other words the
kici%
search above also selects kic
because
the aliases table considers kic
and kici
to
be equivalent. To suppress aliasing use double-quotes
("kic%"
doesn't select the kici
values).
Anchors can be used to force the text to match the
beginning or end of a field. The two anchors are the same as in
various regular expression languages: ^
(carat) to anchor
to the start-of-line; $
(dollar) to anchor to
end-of-line.
Grapheme delimiters of various kinds are provided.
To find only signs joined by hyphens in the transliteration, use a
hyphen, e.g., gu2-e3
. To find only signs separated by a
space, you must use the underline character, e.g., gu2_e3
(space characters between signs default to meaning boolean 'with' as
described in the section on boolean operators). To find signs which
may have either space or hyphen between them, use the tilde,
gu2~e3
.
Field restrictors are provided to restrict the
scope of a search to parts of an ePSD article. Field restrictors
consist of a field-code followed by a colon, and must be attached
directly to the key. To restrict searches only to citation forms,
for example, use c:ba
; to restrict to transliteration,
use t:ba
. The field-codes are the same as the
TOC Matrix codes, but are given in lower case.
Boolean operators are provided. You can search
for keys that co-occur in a record, even when they are different field
types. The and operator requires the right-hand term
to follow the left-hand term within the same field; ab and
ba
selects records only when ba
comes somewhere
after ab
. Spaces between search terms are automatically
taken as and
operators by the search engine, so you can
often simply write, e.g., ba ab
.
The with operator selects records whenever the two
terms co-occur, so ab with ba
selects all the records
where ab
and ba
both occur but in any
order. Because it cannot make sense to consider the order within a
field when the left- and right-hand terms come from different fields,
the search engine automatically converts and
to
with
if the and-operation produces zero results and the
fields of the two operands are different.
The or operator selects records containing either the
left-hand or the right-hand term, e.g., e3 or e11
.
Determinatives can be searched for by putting them
in curly brackets as in the CDLI conventions, e.g.,
{gic}gu-za
. The search engine is designed to find words
regardless of the way determinatives occur in the indexed data:
{gic}gu-za
finds {gic}gu-za
and
gic-gu-za
. In addition, determinatives are ignored when
searching so that, for example, za~en
finds za
{d}en