How to use ICMR-NIN's IFCT 2017 to check what's in your food

India's own food composition tables cover 528 foods and 151 nutrients, measured from samples across six regions. Here is how to turn them into numbers for your own plate.

A home rice plate from Jharkhand with dal and vegetables
Photo: Sureshkhole, CC BY-SA 4.0, via Wikimedia Commons

To use the IFCT, find each ingredient by name or food code in the Indian Food Composition Tables 2017 from ICMR-NIN, read its value per 100 g of raw edible portion, multiply by the raw weight you used, add the ingredients together, and divide by the cooked weight if you want values per 100 g of the dish. Then compare the totals with ICMR-NIN's 2020 RDAs.

Key takeaways
  • 528 foods, 151 components: the scope of IFCT 2017, all measured on Indian samples.
  • Raw, per 100 g: every value except eggs is for the raw edible portion.
  • kJ ÷ 4.18 gives kcal; IFCT prints energy in kilojoules.
  • 2.98×: the weight gain of polished rice on boiling, the key to per-100 g cooked values.
  • 358 kcal and 11.3 g protein in a plate of 60 g rice, 30 g toor dal and 5 g ghee.

A plate of 60 g raw rice, 30 g toor dal and 5 g ghee works out to about 358 kcal, 11.3 g protein and 1.6 mg iron, roughly a fifth of an adult man's daily protein.

Calculated in this post from the ICMR-NIN IFCT 2017 values for rice (A015) and toor dal (B021), and 37 kJ per gram for ghee, in the worked table below, with these example inputs (60 g rice, 30 g dal, 5 g ghee); the protein share is against the 54 g RDA for a 65 kg sedentary man in ICMR-NIN 2020.

The IFCT is the most careful nutrient dataset India has. Its 2017 edition, compiled by T. Longvah, R. Ananthan, K. Bhaskarachary and K. Venkaiah at the National Institute of Nutrition in Hyderabad, reports 151 food components for 528 key foods.

Nothing in it is borrowed from foreign tables: every value comes from Indian food bought at 107 sampling sites, composited in six regional centres and analysed at NIN. That makes it better for Indian food than any app built on US or UK data, as long as you know its three quirks:

  • Almost everything is raw.
  • Energy is in kilojoules.
  • A blank means "below detection", not "no data".
528key foods in IFCT 2017
151food components per food, from protein to individual polyphenols
6regional composites per food: North, East, South, West, Central, North East
2.98×weight gain of polished rice on boiling (Bognár, 2002)

What is the IFCT and where do its numbers come from?

The IFCT 2017 is ICMR-NIN's national food composition table: measured nutrient values for 528 foods that together supply most of what Indians eat, each the mean of composite samples from six regions. It is the successor to tables NIN has published since 1937, when the first Indian table covered about 200 foods.

The story

The IFCT began in a single room. In 1918 Robert McCarrison set up a Beri-Beri Enquiry unit in one laboratory room at the Pasteur Institute in Coonoor, in the Nilgiris. He was invalided to Britain from 1920 to 1922, and in 1923 the enquiry was axed to save money.

It came back in 1925 as the Deficiency Disease Inquiry and grew by 1929 into the Nutrition Research Laboratories, with McCarrison as first director. W. R. Aykroyd took over in 1935, and V. N. Patwardhan became the first Indian director in 1947. The laboratories moved to Hyderabad in 1958 and were renamed the National Institute of Nutrition in 1969, their golden jubilee year. Almost a century after that one room, T. Longvah, who later co-wrote the institute's centenary history, was first author of IFCT 2017.

Longvah T. and Rajendran A.: Commemorating NIN@100, Maternal & Child Nutrition (2020)

The sampling is the reason to trust it. Foods were bought at 107 sites chosen with probability proportional to size, then pooled at regional centres in Delhi (23 sites), Kolkata (20), Bengaluru (19), Mumbai (12), Varanasi (22) and Guwahati (11).

Rice and wheat samples were analysed individually because they are the main staples; other foods were composited, which cuts noise. Each number is printed with a ± standard deviation across regions, so you can see how much a food varies. Rice, for example, is 7.94 ± 0.58 g protein per 100 g.

Code letterFood groupFoods
ACereals and millets24
BGrain legumes25
CGreen leafy vegetables34
DOther vegetables78
EFruits68
GCondiments and spices33
LMilk and milk products4
OAnimal meat63
PMarine fish92
TEdible oils and fats14

The full list has 20 groups from A to T. Every food has a four-character code, a letter for the group and a number for its place, so A015 is milled raw rice and B021 is toor (red gram) dal. Use the code, not the name, when you note values down; "dal" alone could be any of a dozen rows.

Green pigeon pea pods on the plant in a field in Raichur, Karnataka
Pigeon pea in Raichur. IFCT lists it three ways: dal (B021), whole dry (B022) and tender fresh (D067), with very different numbers.Photo: Nanditha Gogate, WELL Labs, CC BY-SA 4.0, via Wikimedia Commons

How do you read an IFCT row?

An IFCT row gives the food code, name, number of regions sampled, and then nutrient values per 100 g edible portion, each as a mean ± standard deviation. Table 1, the one most people need, lists moisture, protein, ash, total fat, dietary fibre (total, insoluble, soluble), available carbohydrate and energy in kJ.

CodeFood (raw)Moisture gProtein gFat gFibre gAvail. carb gEnergy kJEnergy kcalIron mgCalcium mg
A015Rice, raw, milled9.937.940.522.8178.2414913570.657.49
A019Wheat flour, atta11.1010.571.5311.3664.1713403214.1030.94
B021Red gram (toor) dal9.2021.701.569.0655.2313843313.9071.73
B010Green gram (moong) dal9.7723.881.359.3752.5913633263.9343.13
B013Lentil (masoor) dal9.7124.350.7510.4352.531349323n/an/a
L002Milk, whole, cow86.643.264.48n/a4.9430573n/an/a

kcal values are the table's kJ divided by 4.18, the conversion IFCT itself gives. Iron and calcium come from Table 5. Cells marked n/a are values I did not extract for this guide, not gaps in the IFCT.

Four reading rules save most mistakes:

  1. Everything is raw except eggs. The preface says so plainly. 100 g of rice in the table is dry grain, not a bowl of cooked rice.
  2. Energy is in kilojoules. Divide by 4.18 for kcal. IFCT computes energy with factors of 17 kJ per gram for protein and available carbohydrate, 37 for fat and 8 for fibre.
  3. A blank cell means below the detection limit. In PDF copies, blanks also shift the remaining numbers left, so count columns against the header before you copy a value. In the minerals table, rice has nine numbers for ten columns because arsenic was below detection.
  4. "Available carbohydrate" is not "carbohydrate by difference". IFCT sums measured starch and sugars; older tables subtracted everything else from 100, which counts fibre as carbohydrate.
From the table to your plate 1 Find codeA015 riceB021 toor dal 2 Read per100 g rawkJ ÷ 4.18 3 Scale byraw grams× g / 100 4 Add up,÷ cooked wtfor per 100 g 5 Comparewith RDA2020 Weigh ingredients raw. Weigh the finished pot once to get its cooked weight. Vitamins drop with cooking; minerals, protein and energy mostly stay in the pot.
The method in five steps. Weighing raw avoids the biggest error, which is applying raw values to cooked food.

How do you calculate the nutrients in a home recipe?

Multiply each ingredient's per-100 g value by its raw weight divided by 100, then add the results. That gives the nutrients in the whole pot; divide by the number of servings for a portion.

Here is the worked example for one plate of plain dal-rice: 60 g raw milled rice (A015), 30 g raw toor dal (B021) and 5 g ghee. Ghee has no proximate row in Table 1 (it appears only in the fatty-acid tables), so treat it as pure fat at IFCT's 37 kJ per gram.

IngredientRaw gEnergy kJProtein gFat gFibre gIron mgCalcium mg
Rice, milled (A015)60894.64.760.311.690.394.49
Toor dal (B021)30415.26.510.472.721.1721.52
Ghee5185.005.00000
Plate total951494.8 (358 kcal)11.275.784.411.5626.01

Against ICMR-NIN's 2020 numbers for a 65 kg sedentary man (2,110 kcal, 54 g protein RDA, 19 mg iron RDA, 1,000 mg calcium), this plate supplies 17% of energy, 21% of protein, 8% of iron and under 3% of calcium. For a 55 kg sedentary woman (1,660 kcal, 45.7 g protein, 29 mg iron) the shares are 22%, 25% and 5%.

That is the honest picture of plain dal-rice: decent protein, very little iron and almost no calcium. A cup of curd or a leafy sabzi changes the last two columns far more than extra dal.

Use this to check what you just read. Change the raw weights for your own plate; the defaults reproduce the worked example.

Dal-rice plate from IFCT 2017 values

How do you convert raw values to cooked food?

Cooking dilutes; it does not delete. The protein in the pot is the protein you weighed raw.The rule behind yield factors

Boiling mostly adds water, so the nutrients in the pot stay close to the raw totals while the weight rises; to get values per 100 g cooked, divide the raw totals by the cooked weight and multiply by 100. The cleanest way is to weigh the finished pot. If you cannot, use a published yield factor.

The standard reference is Antal Bognár's 2002 tables for the German Federal Research Centre for Nutrition, which list weight-yield factors for about 700 foods. Boiled polished long-grain rice comes out at 2.98 times its raw weight (from 140 measurements), parboiled rice at 2.90, unpolished rice at 2.60, and boiled lentils at 2.73. So 60 g of raw rice becomes about 179 g cooked, and IFCT's 7.94 g protein per 100 g raw becomes about 2.7 g per 100 g cooked rice. Nothing was lost; it was diluted.

Cooking dilutes; it does not delete 100 g raw rice7.94 g protein + water, boil 298 g cooked (×2.98)still 7.94 g protein per 100 g cooked2.66 g protein
Yield factor from Bognár (2002); protein from IFCT 2017 code A015. Water-soluble vitamins are the exception: some leach into cooking water or break down with heat.

Vitamins are the real cooking losses. The Indian Nutrient Databank team, applying USDA retention factors, found calcium and zinc retention of 75 to 100% but vitamin C anywhere from 20 to 100%, with vitamin C, potassium and phosphorus losing the most. If you drain rice water or discard dal soaking water, B vitamins and potassium go with it.

Indian Food Composition Tables 2017Longvah T., Ananthan R., Bhaskarachary K., Venkaiah K. · ICMR-National Institute of Nutrition, Hyderabad · 2017

The national table reports 151 components for 528 foods in 12 tables, from proximates and minerals to amino acids, organic acids, polyphenols and phytate. Samples came from 107 sites grouped into six regions; staples were analysed sample by sample and other foods as regional composites, with each value given as a mean ± SD across regions. Energy uses metabolisable factors of 17, 37, 17 and 8 kJ per gram for protein, fat, available carbohydrate and fibre. Limit: all foods except eggs are raw, there are no cooked dishes, and packaged foods are not covered.

ICMR-NIN website
Development of an Indian Food Composition DatabaseVijayakumar A., Dubasi H.B., Awasthi A., Jaacks L.M. · Current Developments in Nutrition · 2024

The team built the open Indian Nutrient Databank (INDB): 1,095 raw foods (528 from IFCT 2017, 369 from IFCT 2004, the rest matched to UK and US tables) and 1,014 recipes from two Indian cooking manuals plus 148 food-blog recipes. Recipe nutrients were summed from ingredients; a sensitivity analysis applied USDA retention factors, which matched only 42.62% of ingredients and changed median values only slightly (potassium fell from 153 to 149 mg per 100 g). Vegetable samosa came out at 443 kcal per 100 g against 57 for khichdi. Limit: no yield-factor correction, no packaged foods, no lab check of cooked dishes.

Read the paper (free)

Which dals and grains carry the most protein per 100 g?

Among IFCT 2017 staples, masoor dal has the most protein at 24.35 g per 100 g raw, followed by moong dal at 23.88 g and urad dal at 23.06 g; milled rice has 7.94 g. The chart below shows why the cereal-pulse pairing works: the dal triples the protein density of the plate.

Protein per 100 g raw, IFCT 2017
See the data as a table
Protein (g)
Masoor dal (B013)24.35
Moong dal (B010)23.88
Urad dal (B003)23.06
Toor dal (B021)21.7
Chana dal (B001)21.55
Atta (A019)10.57
Rice, milled (A015)7.94

Source: ICMR-NIN: Indian Food Composition Tables 2017, Table 1

Protein quantity is half the story; quality is the other half. ICMR-NIN's 2020 brief note sets the adult RDA at 0.83 g per kg per day and recommends a cereal-legume-milk ratio of 3 : 1 : 2.5 for good protein quality, noting that people on cereal-heavy diets with low-quality protein need about 1 g per kg.

The ranked list of vegetarian protein sources is in high-protein Indian vegetarian foods, iron sources in iron-rich Indian foods, and how fast the carbohydrate in these plates hits the blood in glycaemic index of Indian staples.

A plate of bhindi sabzi, chapati, dal and rice
A mixed plate: each item gets its own IFCT row and raw weight, then the sums are added.Photo: Wind Hashira, CC BY-SA 4.0, via Wikimedia Commons

How do you compare your totals with ICMR-NIN's RDAs?

Compare daily totals, not single meals, with the ICMR-NIN 2020 values; for planning an individual's diet, NIN itself recommends the Estimated Average Requirement (EAR) and treats the RDA as a safe upper target. The 2020 revision cut adult energy needs by 3 to 8 kcal per kg because Indian basal metabolic rates and activity costs measured lower than the old FAO-based figures.

ICMR-NIN 2020Man, 65 kgWoman, 55 kg
Energy, sedentary2,110 kcal1,660 kcal
Energy, moderate work2,710 kcal2,130 kcal
Energy, heavy work3,470 kcal2,720 kcal
Protein RDA (EAR)54.0 g (42.9 g)45.7 g (36.3 g)
Iron RDA (EAR)19 mg (11 mg)29 mg (15 mg)
Calcium RDA (EAR)1,000 mg (800 mg)1,000 mg (800 mg)
Visible fat (oil, ghee)20–50 g a day by energy need; not more than 27 g on 2,000 kcal

If you weigh, scale and sum for three typical days and compare against EAR, you get a far better picture of your diet than any calorie app can give you. Packaged foods are the gap: they are not in IFCT, so take their numbers from the label panel, using how to read an FSSAI food label.

A plate of dal, rice, vegetable curry and fish
Fish (group P) and meat (group O) have their own IFCT rows; cut, skin and bone change the numbers, so match the description carefully.Photo: Gaurav Dhwaj Khadka, CC BY-SA 4.0, via Wikimedia Commons

Questions people ask

Where can I get the IFCT 2017?

ICMR-NIN sells the printed book and hosts a PDF on its website (nin.res.in); its Nutrify India Now app also uses NIN data. NIN's copyright note restricts reproducing the tables in electronic products without permission, so personal use is fine but building an app from it needs NIN's consent.

Are IFCT values for raw or cooked food?

Raw, for every food except eggs. Weigh ingredients raw, or convert cooked weight back to raw with a yield factor (divide cooked rice weight by about 2.98).

Why does my calorie app show different numbers?

Most apps use US or UK tables, cooked-food entries with unknown recipes, or the older carbohydrate-by-difference method. IFCT measures Indian samples and uses available carbohydrate, so its energy values for pulses and millets can differ noticeably.

How accurate are the numbers for my own food?

They are national averages with a stated spread. Rice protein is 7.94 ± 0.58 g, so any one bag may differ by several percent; variety, soil and storage all matter. For diet planning, that error is small next to errors in portion size.

Does IFCT include packaged or restaurant food?

No. It covers key raw foods only. Use the nutrition panel on packets, and treat restaurant dishes as estimates built from a similar home recipe.

Sources

  1. ICMR-NIN: Indian Food Composition Tables 2017 (Longvah T., Ananthan R., Bhaskarachary K., Venkaiah K.) (2017; all food values, sampling design, energy factors, food groups)
  2. ICMR-NIN: Brief note on Nutrient Requirements for Indians, RDA and EAR 2020 (2020; energy, protein, iron, calcium, fat)
  3. Longvah T., Rajendran A.: Commemorating NIN@100: Empowering the nation through nutrition 1918-2018, Maternal & Child Nutrition 16(S3):e13116 (2020; source for the story)
  4. Wikipedia: Robert McCarrison (Beri-Beri Enquiry 1918, closure 1923, revival 1925) (accessed 2026)
  5. Bognár A.: Tables on weight yield of food and retention factors of food constituents, BFE-R-02-03 (2002; yield factors 2.98, 2.90, 2.60, 2.73)
  6. Vijayakumar A. et al.: Development of an Indian Food Composition Database, Curr Dev Nutr 8(7):103790 (2024)
  7. Down To Earth: ICMR-INDIAB diet study, Nature Medicine (2025)
  8. Hyderabad Mail: ICMR-NIN launches Nutrify India Now 2.0 (2024)
IFCT 2017ICMR-NINfood compositionRDA 2020yield factorIndian Nutrient Databankdal ricekilojoules
Sapta

Sapta (Saptarshi Sarkar) is a mechanical engineer turned marketer in Bengaluru and a solo touring rider. He writes the seven Saptaverse blogs. Work with him or read more of Foodblog.

Keep reading in Foodblog

From the other blogs