The Census Bureau has done it again.

Once more, for the 99th straight year, it has produced a tome that all kinds of people -- in business, libraries, academia, research, government, writing -- are clamoring to get.

The 1,057-page volume, published yesterday, is called the "Statistical Abstract of the United States."

At first glance, you might not think its 1,582 tables and 125 pages of appendices that pile fact upon dry fact and figure upon relentless figure are the stuff that inspire poets to write odes.

But the friendly folks at the Census Bureau see it differently. Why, to them it's almost as lovely as that imaginary Grecian are that impelled John Keats into a poetic rhapsody proclaiming that "beauty is truth, truth beauty."

The Abstract "is very close to truth," says Helen E. Teir, who heads the 20-member staff that compild the data for the book. "It's as authoritative as statisics can be, and it supports our quest ofor truth."

Bureau spokesman Henry Smith was more graphic. "Oure is the book for thinkers. The World Almance may have all the oh-gee-gosh facts -- from the tallest buildings to the football yardage -- but we have the real measures of American society. That's the difference."

Indeed, the Abstract does have a lot of measures. Often its figures confirm what we already know. For instance what we already know. For instance there are more divorces now than in past years -- 5.1 per 1,000 persons in 1977, compared with 2.2 in 1960 and 3.5 in 1979.

Marriages? There were more in 1977, 10.1 per 1,000, than in but there were fewer than in 1970, when the figure was 10.6.

In 1970 there were norm people in their marriagable 20s, one source noted. As the decade wore on, there were fewer of them, and more people of all ages had taken to living together without marriage.

A particularly revealing statistic shows that for every 1,000 single women aged 20 to 24,264 of them got married in 1960, 220 in 1970 and 133 in 1976.

But for those who insist the institution is not dying, there are these comforting figures -- remarriages increased from 197,000 in 1960 to 393,000 in 1970 to 518,000 in 1976.

"You can keep pursuing an idea from one figure to the next so that in the end, you are so confused, you don't know what you've proved," the source observed.

The Abstract has measures of our wealth. Last year 97.6 percent of all households had television sets. In 1967 93.6 percent had them, so there wasn't much of an increase. But 76 percent of all households and color sets last year, compared with 15 percent that had them in 1967.

The Abstract tells us much about how we live. More of the elderly widowed people are living alone, for instance. More than 80 percent of them did so in 1977, compared with about 72 percent in 1968. Of the 15.5 million people of all ages living alone, nearly two-thirds are women.

The volume also tells how we die. Heart disease is still the biggest killer although fewer men died of ti in 1976 than in 1960 and 1970. The heart death rate for women was also down from 1960 and 1970 but up from that in 1975. Cancer death rates for both sexes have steadily gone up, particularly those for lung cancer.

One poignant measure of women's longevity is a table showing that for every 100 widowers, there were 531 widows in 1977 -- a great increase over the 475 total in 1970.

Reflecting the rising importance of the nation's Hispanic minority, the Abble on household characteristics -- such as age, size, and type of employment -- that lists Hispanic as well as black households separately. It shows that 37.3 percent of Hispanics are in blue-collar jobs, compared with 26.3 percent for blacks and 26.6 percent for Americans generally.

Parts of the Abstract are hard to read. There are primary individuals and secondary individuals. There are weighted averages and observed proportions and probability samples.

But to Census Bureau staffers, who obviously sing their own Ode to the Abstract, the whole thing is sheer poetry. "We love it dearly," says Teir, "and the beauty of it is that it comes out every year."